# QLogix > QLogix (QLogix LLC) makes AI QMS software for life sciences (an eQMS with CAPA, complaints, change and document control, LIMS, ELN, MES and electronic batch records, CMMS, calibration and validation on one audited record store) and is a QMS implementation and integration partner. It selects, configures, validates and connects quality management systems for emerging and mid-size biotech and pharma, CDMOs and CMOs, and 503B outsourcing facilities, and adds AI workflows that draft, trend and flag while qualified people review and approve. Based in Newport Beach, California; on site across Southern California and remote elsewhere. Key facts: - AI is decision support only: it drafts, summarizes, trends and flags. Review, approval and batch disposition stay with qualified personnel, and AI contributions are recorded in the audit trail. - Platform-agnostic: works with the client's existing QMS and its integrations with MES, LIMS and ERP. - Frameworks: 21 CFR Part 11, 21 CFR Parts 210/211, 21 CFR Part 820 (QMSR), EU GMP Annex 11, EU GMP Annex 22 (draft), ICH Q9(R1), ICH Q10, GAMP 5 (2nd ed.), FDA CSA guidance. - Engagements start with a QMS readiness assessment: processes, systems, data flows and backlog, with baseline metrics agreed before anything changes. ## Services - [QMS digitization](https://qlogix.ai/services/qms-digitization) ([Markdown](https://qlogix.ai/services/qms-digitization.md)): QMS digitization is the move from paper or hybrid quality processes to a validated electronic quality management system. QLogix selects the platform that fits your size and product type, configures it to your SOPs, validates it and migrates the records you need to keep. Includes: Deviations, CAPA, Change control, Complaints, Document control, Training, Supplier quality, APR/PQR. - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) ([Markdown](https://qlogix.ai/services/ai-assisted-investigations.md)): An AI-assisted investigation uses AI to prepare the groundwork of a deviation investigation: the problem statement, similar past events, root cause candidates and a draft CAPA plan. The investigator and QA review, edit and approve every part. The AI drafts; it does not conclude. Includes: IS / IS NOT problem statements, Fishbone, 5-Whys, CAPA effectiveness checks. - [Connected quality data](https://qlogix.ai/services/connected-quality-data) ([Markdown](https://qlogix.ai/services/connected-quality-data.md)): Connected quality data means linking the QMS to the plant and lab systems around it (MES, LIMS, ERP and electronic batch records) so that quality events, batch data and lab results can be trended together. It turns a QMS that records everything into one that also shows patterns before they become recurrences. Includes: MES, LIMS, ERP, Batch records. - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) ([Markdown](https://qlogix.ai/services/validation-of-ai-enabled-systems.md)): Validating an AI-enabled system means showing, with documented evidence, that each AI function is fit for its intended use and stays that way. QLogix follows GAMP 5 (Second Edition) and computer software assurance principles: effort is scaled to the risk of the function, and the evidence is written for an inspector to follow. Includes: Intended use, Risk assessment, Test evidence, CSA. - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) ([Markdown](https://qlogix.ai/services/agentic-workflow-automation.md)): Agentic workflow automation uses orchestrated AI agents to carry out routine quality work across your systems: gathering, checking, drafting and chasing. Each agent works inside defined guardrails, can draft but not approve, and hands decisions to your people. Includes: Orchestrator and specialist agents, Least-privilege access, Draft-only writes, Human approval gates, Audit trail. ## Products - [QLogix Agentic QMS](https://qlogix.ai/products/agentic-qms) ([Markdown](https://qlogix.ai/products/agentic-qms.md)): The QLogix Agentic QMS is an electronic quality management system (eQMS) that keeps change control, CAPA and controlled documents on the same audited, signed records as every other QLogix module. - [QLogix Root cause investigation](https://qlogix.ai/products/root-cause-investigation) ([Markdown](https://qlogix.ai/products/root-cause-investigation.md)): QLogix Root cause investigation is AI-enabled root cause analysis software for GMP manufacturers, sold on its own for organizations that already keep their deviations and CAPAs in another QMS. - [QLogix CAPA management](https://qlogix.ai/products/deviation-capa-management) ([Markdown](https://qlogix.ai/products/deviation-capa-management.md)): QLogix CAPA management is CAPA software in the Agentic QMS that takes a corrective and preventive action plan from the problem that started it to the check that shows it worked. - [QLogix Complaint handling](https://qlogix.ai/products/complaint-handling) ([Markdown](https://qlogix.ai/products/complaint-handling.md)): The QLogix complaint handling module is complaint handling software for pharmaceutical, biotech and medical device manufacturers that turns a complaint as received into a structured complaint record with quoted evidence, a route decided by a person and, where needed, a root cause investigation. - [QLogix Batch record review](https://qlogix.ai/products/batch-record-review) ([Markdown](https://qlogix.ai/products/batch-record-review.md)): The QLogix batch record review module is batch record review software that reviews an executed batch record by exception and lists only the entries that do not meet their instruction, limit or good documentation practice. - [QLogix Change management](https://qlogix.ai/products/change-control) ([Markdown](https://qlogix.ai/products/change-control.md)): QLogix change management is GMP change control software that takes a proposed change to a controlled document, a piece of equipment, a facility area or a utility from request to closure on audited, signed records. - [QLogix Document control](https://qlogix.ai/products/document-control) ([Markdown](https://qlogix.ai/products/document-control.md)): QLogix document control is GMP document management software for controlled documents: SOPs, work instructions, policies, forms, specifications, protocols, reports and master batch records, each with its versions, owner and review period. - [QLogix Risk assessment](https://qlogix.ai/products/quality-risk-management-fmea) ([Markdown](https://qlogix.ai/products/quality-risk-management-fmea.md)): The QLogix risk assessment module is quality risk management software that runs each failure mode and effects analysis (FMEA) as one audited, signed record, from the first draft to the acceptance of the residual risk. - [QLogix Regulatory submissions](https://qlogix.ai/products/regulatory-submissions) ([Markdown](https://qlogix.ai/products/regulatory-submissions.md)): The QLogix regulatory submissions module is regulatory submissions software that documents regulatory affairs' assessment of a change and drafts submission text in which every claim quotes its source. - [QLogix LIMS](https://qlogix.ai/products/lims) ([Markdown](https://qlogix.ai/products/lims.md)): The QLogix LIMS is a laboratory information management system that records every in-process, release, stability and raw material result against an approved specification, on instruments in calibration and reference standards in date. - [QLogix Electronic lab notebook](https://qlogix.ai/products/electronic-lab-notebook) ([Markdown](https://qlogix.ai/products/electronic-lab-notebook.md)): The QLogix electronic lab notebook (ELN) is the analyst's contemporaneous record of laboratory work, kept on the same audited, signed records as the LIMS and the quality system. - [QLogix MES (electronic batch records)](https://qlogix.ai/products/mes-electronic-batch-records) ([Markdown](https://qlogix.ai/products/mes-electronic-batch-records.md)): The QLogix MES is a manufacturing execution system that runs electronic batch records from the client's own master batch record, for therapeutic products, medical devices and combination products. - [QLogix ERP](https://qlogix.ai/products/erp-materials-inventory) ([Markdown](https://qlogix.ai/products/erp-materials-inventory.md)): The QLogix ERP runs the material and product flow of a GMP plant on the same audited, signed records as the quality system, the LIMS and the MES. - [QLogix GMP logbooks](https://qlogix.ai/products/gmp-electronic-logbooks) ([Markdown](https://qlogix.ai/products/gmp-electronic-logbooks.md)): QLogix GMP logbooks are electronic logbooks for cleanroom cleaning, equipment use and QC instruments, kept as audited records on the same platform as the quality system. - [QLogix PM Library](https://qlogix.ai/products/pm-library) ([Markdown](https://qlogix.ai/products/pm-library.md)): The QLogix PM Library is a particulate matter and defect library for parenteral fill-finish: every particle and defect type the filling process has produced, characterized with high-resolution images and its material identity from FTIR, Raman or SEM-EDX, placed where it enters the product relative to the 0.2 µm sterilizing filtration, and approved by QA. - [QLogix Stability Study Management](https://qlogix.ai/products/stability-study-management) ([Markdown](https://qlogix.ai/products/stability-study-management.md)): QLogix Stability Study Management runs a pharmaceutical stability program in one system, on the same audited records as the LIMS and the QMS. - [QLogix Calibration management](https://qlogix.ai/products/calibration-management) ([Markdown](https://qlogix.ai/products/calibration-management.md)): QLogix calibration management keeps one audited record per instrument, equipment, tool, gauge, test equipment, sensor or reference standard under calibration. - [QLogix CMMS](https://qlogix.ai/products/cmms) ([Markdown](https://qlogix.ai/products/cmms.md)): The QLogix CMMS is a computerized maintenance management system built on the same audited, signed records as the quality system. - [QLogix Validation](https://qlogix.ai/products/computer-system-validation) ([Markdown](https://qlogix.ai/products/computer-system-validation.md)): The QLogix validation module is computer system validation software that runs the whole validation lifecycle on audited, signed records, for computerized systems and for equipment, facilities and utilities, processes and cleaning. - [QLogix Project management](https://qlogix.ai/products/project-management) ([Markdown](https://qlogix.ai/products/project-management.md)): QLogix project management runs programs and projects on the same audited records as the quality system. - [QLogix Operational excellence](https://qlogix.ai/products/operational-excellence) ([Markdown](https://qlogix.ai/products/operational-excellence.md)): QLogix operational excellence runs continuous improvement on the same audited, signed records as the quality system. - [QLogix Q Intelligence, Q Agent and agents](https://qlogix.ai/products/ai-for-pharma-quality) ([Markdown](https://qlogix.ai/products/ai-for-pharma-quality.md)): QLogix Q Intelligence and Q Agent are AI for pharmaceutical quality: an AI quality assistant and AI agents that work on the same audited records as the quality system, while qualified people decide and sign. ## White papers - [From deviation to defensible root cause: AI-enabled investigations for GxP manufacturing](https://qlogix.ai/whitepapers#ai-enabled-investigations) (2026-10-15; one-page PDF, free on request): Why inadequate investigations remain among FDA's most frequent observations, and how an AI-drafted analysis, hypothesis testing and an 11-point root cause check keep every decision with the investigator. Product: https://qlogix.ai/products/deviation-capa-management - [One quality system, one AI layer: The case for an Agentic QMS](https://qlogix.ai/whitepapers#agentic-qms) (2026-10-15; one-page PDF, free on request): What separate quality systems cost in validation, interfaces and oversight, and how one compliant record layer lets AI agents see the whole quality system and act only with a person's approval. Product: https://qlogix.ai/products/agentic-qms - [Risk assessments that stay true to the data: Q Intelligence and the AI-enabled FMEA](https://qlogix.ai/whitepapers#quality-risk-management-fmea) (2026-10-15; one-page PDF, free on request): Quality risk management under ICH Q9(R1): occurrence and detection scored from the organization's own records, every Critical and High risk followed up, and the residual risk accepted on evidence. Product: https://qlogix.ai/products/quality-risk-management-fmea - [From change control to filed dossier: AI-enabled regulatory submissions on an Agentic QMS](https://qlogix.ai/whitepapers#regulatory-submissions) (2026-10-15; one-page PDF, free on request): Post-approval changes screened for five markets, NDA and BLA dossiers checked against a 62-section eCTD template, and submission text in which every claim quotes its source. Product: https://qlogix.ai/products/regulatory-submissions - [From complaint to closed CAPA, with nothing left behind: An Agentic QMS for complaints, investigations and backlog](https://qlogix.ai/whitepapers#complaints-investigations-backlog) (2026-10-15; one-page PDF, free on request): The reporting clocks a complaint can start, how AI triage hands each complaint to an investigation with its evidence, and how a risk-ranked backlog puts the oldest serious records first. Product: https://qlogix.ai/products/complaint-handling - [Lab and plant, connected: LIMS, calibration, logbooks and CMMS on one AI-enabled infrastructure](https://qlogix.ai/whitepapers#connected-lab-and-plant) (2026-10-15; one-page PDF, free on request): How shared records turn GMP dependencies into rules: no result on an out-of-calibration instrument, calibration failures feeding maintenance reliability, and logbook reviews the system checks first. Product: https://qlogix.ai/products/lims - [Validation, digitized and on schedule: Electronic validation with project management built in](https://qlogix.ai/whitepapers#validation-and-project-management) (2026-10-15; one-page PDF, free on request): Eighteen templated validation documents on the V-model, electronic test execution with traceability, and a schedule that reports every day whether the validation project is on track. Product: https://qlogix.ai/products/computer-system-validation - [From particle to qualified inspection: An AI-enabled particulate matter library for fill-finish](https://qlogix.ai/whitepapers#pm-library-fill-finish) (2026-10-15; one-page PDF, free on request): Visual inspection is only as good as the defects people and machines are trained on. How a characterized, versioned library keeps inspector requalification and AVI challenge sets current with the particles the process actually produces. Product: https://qlogix.ai/products/pm-library ## Frequently asked questions ### About QLogix - **What does QLogix do?** QLogix provides AI transformation for life sciences quality and operations in two ways: an AI applications platform for GxP operations (eQMS, root cause investigation, CAPA, complaints, change and document control, risk, regulatory submissions, LIMS, ELN, stability, MES, ERP, logbooks, calibration, CMMS, validation, project management and operational excellence), and implementation services that select, configure, validate and connect quality systems. AI drafts, trends and flags; qualified people review, decide and sign. - **Who is QLogix for?** QLogix works with emerging and mid-size biotech and pharma companies moving from paper or hybrid systems to validated, connected ones; CDMOs and CMOs running several client quality agreements on one system; 503B outsourcing facilities meeting cGMP expectations with lean teams and high batch volume; and medical device and combination product manufacturers. - **What is an AI QMS?** An AI QMS is a quality management system in which AI drafts records, investigations and assessments from the evidence on the record, quotes its sources and routes every decision to a qualified person, who approves with an electronic signature. In the QLogix Agentic QMS every AI contribution is an entry in the same hash-chained audit trail as the people's. ### The platform - **Which products does the QLogix platform include?** Twenty-two modules in five groups. Quality events: root cause investigation, CAPA management, complaint handling and batch record review. Documents, change and risk: change management, document control, risk assessment (FMEA) and regulatory submissions. Laboratory, manufacturing and equipment: LIMS, electronic lab notebook, stability study management, MES with electronic batch records, ERP, GMP logbooks, the PM Library, calibration management and CMMS. Validation, projects and improvement: computer system validation, project management and operational excellence. AI for quality: Q Intelligence, Q Agent and agents. The Agentic QMS ties them together. - **Are the modules connected?** Yes. Every QLogix module keeps its records on one audited record store, so a deviation links to the batch in the MES, the result in the LIMS, the work order in the CMMS and the CAPA that closes it, in both directions. Each record has a SHA-256 hash-chained audit trail, review and approval with electronic signatures, attachments and links, and each organization sees only its own data. - **Can we buy a single module?** Yes. Several QLogix modules are sold on their own: for example root cause investigation for organizations that keep deviations and CAPAs in another QMS, or the CMMS, ELN, calibration management, project management, ERP or operational excellence. They connect to the rest of the platform when you add more. - **Does QLogix connect to our other systems?** Yes. QLogix connects to ERP, LIMS, MES, BI tools, data warehouses, integration platforms such as Power Automate, Zapier, Workato or MuleSoft, and AI assistants, through a REST API and an MCP server. Keys are issued per client workspace on request and read that workspace's data only. ### Quality events - **How does root cause investigation work?** QLogix Root cause investigation takes a deviation, OOS or OOT result, complaint or audit finding from the first write-up to a tested root cause and a corrective action plan. The AI drafts the problem statement with IS / IS NOT, the 5 Whys, the fishbone and a CAPA proposal from the write-up and documents; the investigator searches related records, builds the sequence of events and the change analysis, tests each cause as a hypothesis, scores the risk (ICH Q9) and the cost of poor quality, and decides. It works on its own or with the Agentic QMS. - **How is CAPA effectiveness shown?** In QLogix CAPA management each corrective and preventive action has an owner, a due date and, when done, its evidence, and is rated on the hierarchy of controls so retraining alone is flagged. Effectiveness is measured from data: the counts before and after the actions give a one-sided exact test and an upper confidence bound, with the residual risk and a business case for the CAPA alongside. - **What do complaint handling and batch record review do?** QLogix complaint handling turns a complaint as received into a structured record with quoted evidence, a route decided by a person and, where needed, a root cause investigation. QLogix batch record review reviews an executed batch record by exception, electronic or scanned paper, and lists only the entries that miss their instruction, limit or good documentation practice, for QA to disposition. ### Documents, change and risk - **How do change control and document control work?** QLogix change management takes a change to a document, piece of equipment, area or utility from request to closure: the AI assesses its impact, risk and the validation it needs as soon as it is raised, and the functions affected sign off. QLogix document control manages SOPs, work instructions, specifications and master batch records with their versions and review periods; the AI reviews each draft, and approved versions become effective with read-and-understood training. - **How are risk assessments and regulatory submissions handled?** The QLogix risk assessment module runs each design, process, equipment or CSV FMEA as one signed record, from the AI's first draft of the register to the acceptance of the residual risk, and investigations feed new failure modes back into it. The regulatory submissions module screens a change for regulatory impact and drafts submission text in which every claim quotes its source; regulatory affairs decides what is filed. ### Laboratory, manufacturing and equipment - **What do the LIMS, ELN and stability modules cover?** The QLogix LIMS records in-process, release, stability and raw material results against approved specifications, calculates each result from the determinations and judges it against the limits, on instruments in calibration and standards in date. The electronic lab notebook is the analyst's contemporaneous record of runs, preparations and method work. Stability study management builds protocols on ICH Q1A(R2) conditions, schedules the pulls and trends the results. - **What do the MES and ERP do?** The QLogix MES runs electronic batch records from your own master batch record, drafted by the AI from your existing record and approved by QA, for therapeutic products, medical devices and combination products. The QLogix ERP runs materials and products from the item master and purchasing through receipt, QC release, lot genealogy and inventory, on the same records as quality, the LIMS and the MES. - **What about equipment, logbooks and visual inspection?** QLogix calibration management keeps one record per instrument or standard with its approved requirement and each calibration's result. The QLogix CMMS manages assets with their criticality, GxP impact and preventive maintenance, and work orders with labour, parts and downtime. GMP logbooks record cleaning, equipment use and instrument entries contemporaneously with second-person verification. The PM Library characterizes the particles and defects of parenteral fill-finish for visual inspection. ### Validation, projects and improvement - **How does computer system validation work in QLogix?** The QLogix validation module runs the validation lifecycle on signed records for computerized systems, equipment, facilities and utilities, processes and cleaning: an AI-drafted validation plan, risk-based requirements and test scripts, executed tests with deviations, a traceability matrix and the summary report, following GAMP 5 (Second Edition) and computer software assurance. - **What do project management and operational excellence add?** QLogix project management runs programs and projects with an AI-drafted plan, resources and load, risks, the critical path and earned value. QLogix operational excellence runs continuous improvement tools as applications: A3, kaizen, value stream maps, 5S, standard work, daily management, KPI bowlers and policy deployment, with computed results and action plans. A CAPA action can start the improvement tool that fits it, linked back. ### AI for quality - **What are Q Intelligence and Q Agent?** QLogix Q Intelligence is an AI quality assistant that answers questions about any record, document, audit trail or regulation, shows every source it consulted and checks its citations; every answer is logged for validation. QLogix Q Agent carries out multi-step work across modules, and every action it proposes waits for a person's approval. - **Does the AI make quality decisions?** No. In QLogix the AI drafts, summarizes, trends and flags. Review, approval, root cause conclusions, risk acceptance and batch disposition stay with qualified people, who sign electronically, and each AI contribution is recorded in the audit trail with the model and the sources it used. ### Services - **Which services does QLogix offer?** Five services: QMS digitization (selecting, configuring, validating and migrating to an electronic QMS); AI-assisted investigations (problem statements, root cause candidates and CAPA plans drafted for investigators to approve); connected quality data (linking the QMS to MES, LIMS, ERP and batch records for trending); validation of AI-enabled systems (risk-based, to GAMP 5 and CSA); and agentic workflow automation (AI agents that gather, check, draft and chase within guardrails, with human approval). - **Do we have to replace our QMS?** Usually not. QLogix is platform-agnostic: it works with your current QMS and its integrations with MES, LIMS and ERP, and recommends a change of platform only when the assessment shows the current one cannot meet the need. Modules such as root cause investigation can sit beside an existing QMS. - **How are AI functions validated?** QLogix validates AI functions with a risk-based approach: defined intended use, a risk assessment, documented testing against acceptance criteria, monitoring in use, and change control for any model or prompt update. Q Intelligence comes with an intended use, a risk assessment, an automated controls check and an append-only answer log as validation evidence. ### Compliance - **Which regulations and frameworks does QLogix work to?** QLogix works to 21 CFR Part 11, 21 CFR Parts 210 and 211, 21 CFR Part 820 (QMSR), EU GMP Annex 11, the draft EU GMP Annex 22 on AI, ICH Q9(R1), ICH Q10, ICH Q1A(R2) for stability, GAMP 5 (Second Edition) and FDA's computer software assurance guidance, with ALCOA+ data integrity throughout. - **How does QLogix meet Part 11 and data integrity expectations?** Every QLogix record has an append-only audit trail chained with SHA-256 for tamper evidence, electronic signatures with re-authentication and the meaning of the signature, role-based access with review and approval separated by default, and attributable, time-stamped entries. Records are corrected with a reason, never deleted, and exports of the controlled copy are logged. ### Working with QLogix - **How does an engagement start?** Most QLogix clients start with a QMS readiness assessment covering processes, systems, data flows and backlog. Baseline metrics such as deviation cycle time, backlog age, recurrence and right first time are agreed before anything changes, so the improvement can be measured. - **Where does QLogix work?** QLogix is based in Newport Beach, California, and works on site across Southern California, with remote delivery elsewhere in the United States. ## Blog - [Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviation](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) ([Markdown](https://qlogix.ai/blog/spc-signal-detection-gmp-trending.md); Manufacturing science, updated 2026-10-15): A result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification. - [Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by age](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) ([Markdown](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization.md); AI in quality, updated 2026-10-15): Most quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible. - [Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replaced](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ([Markdown](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp.md); AI in quality, updated 2026-10-15): A preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control. - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens compliance](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ([Markdown](https://qlogix.ai/blog/lims-investigation-capa-integration-ai.md); AI in quality, updated 2026-10-14): Most OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign. - [AI in deviation investigations: faster, more consistent root cause analysis](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) ([Markdown](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis.md); AI in quality, updated 2026-10-14): AI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable hypotheses, each quoting the write-up or an attached document. The investigator then tests, decides and signs. Nothing the AI writes becomes the root cause until a qualified person says so. - [AI-assisted CAPA: from root cause to a verified effectiveness check](https://qlogix.ai/blog/ai-capa-management-effectiveness) ([Markdown](https://qlogix.ai/blog/ai-capa-management-effectiveness.md); AI in quality, updated 2026-10-14): A CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same cause coming back. AI helps most before the plan is approved: proposing actions that address the confirmed cause, drafting the effectiveness criterion up front, and matching the CAPA against earlier records. People approve, implement, verify and close. - [AI in document management and validation workflows for life sciences](https://qlogix.ai/blog/ai-document-management-validation-workflow) ([Markdown](https://qlogix.ai/blog/ai-document-management-validation-workflow.md); Validation, updated 2026-10-14): Document control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. The same holds for validation: plans, protocols and reports drafted from templates and requirements, with assurance effort set by risk under GAMP 5 and FDA's CSA guidance, and a person approving every version with an electronic signature. - [FDA's seven-step credibility framework for AI, in plain terms](https://qlogix.ai/blog/fda-ai-credibility-framework) ([Markdown](https://qlogix.ai/blog/fda-ai-credibility-framework.md); AI in quality, updated 2026-10-12): FDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the decision, and how serious a wrong decision would be. - [AI in drug manufacturing: where FDA stands and what to build for](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) ([Markdown](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda.md); Manufacturing science, updated 2026-10-12): FDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing models. Together they show what a process model has to prove. - [AI-enabled medical devices: what FDA's list and guidance show](https://qlogix.ai/blog/ai-enabled-medical-devices-fda) ([Markdown](https://qlogix.ai/blog/ai-enabled-medical-devices-fda.md); Medical devices, updated 2026-10-12): FDA's public list counted 1,614 AI-enabled medical devices through June 2026, about three in four of them in radiology. Two documents shape how they are built and changed: the final guidance on predetermined change control plans and the draft guidance on lifecycle management. - [Robotics in aseptic manufacturing: what Annex 1 expects](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1) ([Markdown](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1.md); Manufacturing science, updated 2026-10-12): The revised EU GMP Annex 1 names robotic systems, with isolators and RABS, as technologies to consider for protecting sterile products. Taking the operator out of the critical zone removes the largest contamination source and moves the quality work to design, software and data. - [EU GMP Annex 22: what the draft AI annex asks of GMP manufacturers](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft) ([Markdown](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft.md); Regulatory updates, updated 2026-10-12): Draft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined intended use, independent test data and ongoing monitoring. - [Computer software assurance: what FDA's final guidance changes for validation](https://qlogix.ai/blog/computer-software-assurance-final-guidance) ([Markdown](https://qlogix.ai/blog/computer-software-assurance-final-guidance.md); Validation, updated 2026-10-12): FDA's final guidance on computer software assurance replaces test-everything validation with effort matched to risk. Features that could compromise safety get scripted testing; the rest can be covered by unscripted testing and supplier evidence. ## Pages - [Home: QMS implementation for life sciences](https://qlogix.ai/) ([Markdown](https://qlogix.ai/index.md)): AI transformation solutions for life sciences: AI QMS software and implementation for pharma, biotech, CDMOs and 503B, from eQMS and CAPA to LIMS and validation. - [Services](https://qlogix.ai/services) ([Markdown](https://qlogix.ai/services.md)): QLogix services: QMS digitization, AI-assisted investigations, connected quality data, validation of AI-enabled systems and agentic workflow automation. - [QMS digitization](https://qlogix.ai/services/qms-digitization) ([Markdown](https://qlogix.ai/services/qms-digitization.md)): Move deviations, CAPA, change control, complaints, documents and training from paper or hybrid systems to a validated electronic QMS, configured to your SOPs. - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) ([Markdown](https://qlogix.ai/services/ai-assisted-investigations.md)): AI drafts the problem statement, root cause candidates and CAPA plan from your records. Investigators and QA review, edit and approve, with a full audit trail. - [Connected quality data](https://qlogix.ai/services/connected-quality-data) ([Markdown](https://qlogix.ai/services/connected-quality-data.md)): Link your QMS to MES, LIMS, ERP and batch records for trending, recurrence detection and earlier signals, with validated interfaces and data integrity controls. - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) ([Markdown](https://qlogix.ai/services/validation-of-ai-enabled-systems.md)): Risk-based validation of AI functions in GxP systems: intended use, risk assessment, test evidence against acceptance criteria and change control for model updates. - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) ([Markdown](https://qlogix.ai/services/agentic-workflow-automation.md)): AI agents that gather, check, draft and chase across QMS, LIMS, MES and ERP, inside guardrails: least-privilege access, draft-only writes and human approval. - [Products](https://qlogix.ai/products) ([Markdown](https://qlogix.ai/products.md)): QLogix products: eQMS, CAPA and deviations, complaints, change and document control, LIMS, ELN, MES and batch records, CMMS, calibration and validation. - [Agentic QMS](https://qlogix.ai/products/agentic-qms) ([Markdown](https://qlogix.ai/products/agentic-qms.md)): eQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions. - [Root cause investigation](https://qlogix.ai/products/root-cause-investigation) ([Markdown](https://qlogix.ai/products/root-cause-investigation.md)): AI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan. - [CAPA management](https://qlogix.ai/products/deviation-capa-management) ([Markdown](https://qlogix.ai/products/deviation-capa-management.md)): GMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check. - [Complaint handling](https://qlogix.ai/products/complaint-handling) ([Markdown](https://qlogix.ai/products/complaint-handling.md)): Complaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started. - [Batch record review](https://qlogix.ai/products/batch-record-review) ([Markdown](https://qlogix.ai/products/batch-record-review.md)): Batch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation. - [Change management](https://qlogix.ai/products/change-control) ([Markdown](https://qlogix.ai/products/change-control.md)): GMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure. - [Document control](https://qlogix.ai/products/document-control) ([Markdown](https://qlogix.ai/products/document-control.md)): Document control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review. - [Risk assessment](https://qlogix.ai/products/quality-risk-management-fmea) ([Markdown](https://qlogix.ai/products/quality-risk-management-fmea.md)): Quality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk. - [Regulatory submissions](https://qlogix.ai/products/regulatory-submissions) ([Markdown](https://qlogix.ai/products/regulatory-submissions.md)): Regulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source. - [LIMS](https://qlogix.ai/products/lims) ([Markdown](https://qlogix.ai/products/lims.md)): LIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA. - [Electronic lab notebook](https://qlogix.ai/products/electronic-lab-notebook) ([Markdown](https://qlogix.ai/products/electronic-lab-notebook.md)): An ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS. - [MES (electronic batch records)](https://qlogix.ai/products/mes-electronic-batch-records) ([Markdown](https://qlogix.ai/products/mes-electronic-batch-records.md)): MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release. - [ERP](https://qlogix.ai/products/erp-materials-inventory) ([Markdown](https://qlogix.ai/products/erp-materials-inventory.md)): A GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning. - [GMP logbooks](https://qlogix.ai/products/gmp-electronic-logbooks) ([Markdown](https://qlogix.ai/products/gmp-electronic-logbooks.md)): Electronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed. - [PM Library](https://qlogix.ai/products/pm-library) ([Markdown](https://qlogix.ai/products/pm-library.md)): A particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification. - [Stability Study Management](https://qlogix.ai/products/stability-study-management) ([Markdown](https://qlogix.ai/products/stability-study-management.md)): Stability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP. - [Calibration management](https://qlogix.ai/products/calibration-management) ([Markdown](https://qlogix.ai/products/calibration-management.md)): Every instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA. - [CMMS](https://qlogix.ai/products/cmms) ([Markdown](https://qlogix.ai/products/cmms.md)): A GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history. - [Validation](https://qlogix.ai/products/computer-system-validation) ([Markdown](https://qlogix.ai/products/computer-system-validation.md)): Computer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports. - [Project management](https://qlogix.ai/products/project-management) ([Markdown](https://qlogix.ai/products/project-management.md)): Programs and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval. - [Operational excellence](https://qlogix.ai/products/operational-excellence) ([Markdown](https://qlogix.ai/products/operational-excellence.md)): A3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps. - [Q Intelligence, Q Agent and agents](https://qlogix.ai/products/ai-for-pharma-quality) ([Markdown](https://qlogix.ai/products/ai-for-pharma-quality.md)): AI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set. - [Blog](https://qlogix.ai/blog) ([Markdown](https://qlogix.ai/blog.md)): QLogix blog: regulatory updates on AI in GxP, FDA and EU guidance explained, and how life sciences quality teams adopt and validate AI. - [White papers](https://qlogix.ai/whitepapers) ([Markdown](https://qlogix.ai/whitepapers.md)): QLogix white papers: AI investigations, agentic QMS, FMEA, regulatory submissions, complaints, LIMS and CMMS, validation and particulate inspection. - [Request a meeting](https://qlogix.ai/contact) ([Markdown](https://qlogix.ai/contact.md)): Talk to QLogix about QMS implementation, AI-assisted investigations and CAPAs, connected quality data and validation of AI-enabled systems. - [Site map](https://qlogix.ai/sitemap) ([Markdown](https://qlogix.ai/sitemap.md)): Every public QLogix page: services, AI QMS products, blog posts and white papers, each with its Markdown version for AI assistants. - [Terms of Use](https://qlogix.ai/termsofuse) ([Markdown](https://qlogix.ai/termsofuse.md)): The legal terms for using the QLogix website and services: accounts, acceptable use, intellectual property, disclaimers, liability and dispute resolution. - [Privacy Policy](https://qlogix.ai/privacy) ([Markdown](https://qlogix.ai/privacy.md)): How QLogix collects, uses, shares and protects personal information, the cookies it uses, and the privacy rights and choices you have. ## For language models - [Full text of every public page](https://qlogix.ai/llms-full.txt) - Every page is also available as Markdown at its URL plus ".md", for example https://qlogix.ai/services/qms-digitization.md (the home page is https://qlogix.ai/index.md). Each Markdown page starts with its title, description and canonical URL, and links to the other Markdown pages. - Requesting any page with the header "Accept: text/markdown" returns its Markdown at the same address. - [Site map](https://qlogix.ai/sitemap) and [XML site map](https://qlogix.ai/sitemap.xml). - When citing QLogix, link the page's canonical URL (without ".md"). ## Contact - Email: contact@qlogix.ai - Phone: +17146639012 - Address: 5000 Birch Street, West Tower, Suite 3000, Newport Beach, CA 92660 --- # QLogix - AI Applications Platform for all GxP Operations Source: https://qlogix.ai/ One AI Platform for Every GxP Operation. # Close deviations and CAPAs faster, without loosening control. QLogix delivers AI transformation for life sciences quality. We select, configure, validate and connect quality management systems for biotech, pharma, CDMOs and 503B facilities, then add AI workflows that draft, trend and flag, while qualified people review and approve. [Request a meeting](https://qlogix.ai/contact)[See it on a deviation](#walkthrough) Designed around - 21 CFR Part 11 - EU GMP Annex 11 - GAMP 5 (2nd ed.) - ICH Q10 Platform-agnostic. We implement and integrate with your existing quality management system. Custom-builtto your site's specific requirements: the workflows, records and integrations you need and no product has. ## Quality systems that record everything and connect nothing Most sites we see run a capable QMS next to MES, LIMS and ERP systems that never talk to it. The cost shows up in the same places every time. ### Investigations that start from a blank page Each deviation is written up from scratch, so quality varies by author and cycle time depends on who is available. [Try the simulation](#sim-rootcause) ### Backlogs that age quietly Open deviations, CAPAs and change controls pile up without a clear view of which ones carry risk. [Try the simulation](#sim-backlog) ### Signals found after the batch Trends sit across batch records, lab results and complaints. Nobody sees the pattern until it becomes a recurrence. [Try the simulation](#sim-signals) ## See it work Nine short simulations with illustrative data. They start on their own as you scroll: the AI does the groundwork, the architecture above each one shows where the work happens, and every decision waits for you. Root cause investigation Backlog triage Signal detection Complaint handling Batch record review Regulatory submissions Inspection readiness Supplier risk Validation planning ### From a blank page to an approved CAPA Tour playingIllustrative data Concept architecture Connected systems - QMS - MES - LIMS - DMS One record Hash-chained audit trail, verified on every read AI drafts Drafts six root cause candidates, then the CAPA People decide the investigator and QA Starting… - audit trail - empty until the simulation starts Draft root cause candidates AI drafts one candidate per category from the evidence. Method Machine Material DEV-0139 · problem Particulates found at visual inspection Sterile fill, line 2 Measurement People Environment The AI drafts candidates and the CAPA. Your investigator and QA make every decision. Want to see this with your own systems and data? [Request a meeting](https://qlogix.ai/contact) ## Run the numbers on your own site Five models you can drive. Move the sliders, scrub the timeline, break the chain: the charts answer with the arithmetic the platform runs on real records. Backlog simulatorMonte Carlo, 200 runs Project forecastEarned value Risk heatmapFMEA Audit trail chain21 CFR Part 11 Systems mapArchitecture ### What happens to your deviation backlog over a year? Set your site's intake, team and cycle time. The model runs the year two hundred times and shows where the backlog lands with and without the platform. Illustrative model New deviations per month30 Investigators5 Average investigation, open to close35 days Deviations that recur20 % With QLogix AI drafts the investigationabout 35 % shorter cycle, open to close LIMS and MES data connectedtrends caught early: about 50 % fewer repeats Open after 12 months 25 109 fewer than without Closed per month 33 26 without Repeat deviations, year 40 61 without Investigation cycle 23 d 35 d without With QLogix, median80% of runsWithout, median Each investigator carries about 6 open investigations at a time, so the team closes roughly 26 a month at a 35 -day cycle. Arrivals vary month to month and 20 % of closed deviations come back. Drag the sliders to your own site; the assessment measures the real figures. The assessment runs these models on your real records. [Book an assessment](https://qlogix.ai/contact) ## Agents that take on the routine work Agentic automation handles the gathering, checking, drafting and chasing that fills a quality team's day. Agents work across your systems inside guardrails you define, so routine tasks get done faster, records stay inspection-ready, and your people spend their time on decisions. Agentic architecture Illustrative Orchestrator agent: plans, routes and assembles Intake agent - Complaint inbox - Call center - Web forms Investigation agent - QMS - Batch records Trending agent - LIMS - MES - QMS Batch review agent - eBR - MES Document agent - DMS - Regulatory files Guardrails - Least-privilege access per agent - Draft-only writes - Human approval gates - Every tool call in the audit trail - Models and prompts under change control The day's scheduled tasks Scheduled runs, shown complete 06:00Prepare the morning quality huddle06:15Chase overdue CAPA actions06:30Assemble the APR/PQR data pack - 06:00:09 [ orchestrator ] plan: new events, records due this week, batches awaiting review, complaints - 06:00:18 [ intake ] read 4 new complaints, 1 flagged as a possible adverse event - 06:00:21 [ investigation ] queried QMS: 3 deviations due in 7 days, 1 overdue - 06:00:30 [ batch ] checked the eBR queue: 2 batches ready, 1 with exceptions - 06:00:39 [ trending ] reviewed EM and in-process results: no new trends - 06:00:42 [ orchestrator ] drafted the huddle brief: 3 items need a decision Ready for your review - DEV-0142 impact assessment overdue: assign an owner - Call center complaint flagged as a possible adverse event: route to pharmacovigilance - Lot 24-121 batch record has 3 exceptions: assign a reviewer Approve the brief : approved by the QA lead. Logged with every source and agent step. Shown complete. Each task ends at a human approval. ### Why an agentic architecture - Routine work runs on schedule Huddle prep, overdue chasing and review packs are ready when your team starts the day. - One task, many systems An orchestrator splits the work across agents that read QMS, LIMS, MES, ERP and documents in parallel. - Bounded by design Each agent has a defined intended use, only the access it needs, and can draft but not approve. - Inspection-ready by default Sources, tool calls and drafts are recorded, so an inspector can follow how every output was produced. - People stay accountable Approvals, dispositions and reportability calls stay with qualified people. Agents bring them the work. Which routine tasks would you hand to agents first? [Request a meeting](https://qlogix.ai/contact) ## One deviation, start to close Here is where AI does the groundwork and where your people make the call, at each stage of a deviation. 01Event02Investigation03Root cause04CAPA05Effectiveness ### Deviation logged AI drafts and flags - Drafts the problem statement in IS / IS NOT form from the initial report - Flags similar past events across batches, equipment and sites Your qualified people decide - Confirm the facts and the initial impact assessment - Decide the classification and whether product is affected audit trail · 14:02 draft generated · 14:10 routed to QA reviewer Want to see how this would work with your systems and SOPs? [Request a meeting](https://qlogix.ai/contact) ## One platform, every module on the same audited records Licence the modules you need, or all of them, per user. Every module writes to the same hash-chained audit trail, signs with the same electronic signatures and links to every other record, so an OOS result, its investigation, the CAPA, the change control and the calibration of the instrument read as one story. ### Quality events - [Complaint handling Triage with the adverse-event check, evidence quoted from the complaint, and the decision signed.](https://qlogix.ai/products/complaint-handling) - [Root cause investigation Deviation, OOS and complaint investigations with AI-drafted problem statements, hypotheses tested and a robust root cause; on its own beside your existing QMS.](https://qlogix.ai/products/root-cause-investigation) - [CAPA management CAPAs opened from investigations, complaints, audits and changes, with an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Batch record review Exceptions found, assessed and dispositioned, with the recommendation signed by QA.](https://qlogix.ai/products/batch-record-review) - [Signal detection Control charts with run rules on any parameter; each signal interpreted by the AI and dispositioned by QA.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [Backlog triage Open deviations, CAPAs and change controls ranked by risk, with QA's priorities on record.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) ### Documents and change - [Change management Changes to documents, equipment, facilities and utilities under GMP: what changes and how, AI impact and risk assessment, the GMP requirements each item must meet, tasks, approvals to implement, effectiveness and closure.](https://qlogix.ai/products/change-control) - [Document management Controlled documents with versions, AI review, approvals, training, effective dates and periodic review; every QMS document in one index.](https://qlogix.ai/products/document-control) - [Regulatory submissions Changes screened for regulatory impact; NDA and BLA sections drafted from templates.](https://qlogix.ai/products/regulatory-submissions) - [Risk assessment Design, process, equipment and CSV FMEA with RPN thresholds, mitigation actions and reassessment.](https://qlogix.ai/products/quality-risk-management-fmea) ### Laboratory, equipment and maintenance - [LIMS Specifications, samples and results judged by the system, OOS handled under investigation, instruments and standards locked out when out of date.](https://qlogix.ai/products/lims) - [Electronic lab notebook Each analytical run recorded as it happens: the effective method, equipment checked across the LIMS, calibration and maintenance, standards and reagents in date, observations never edited, results posted to the LIMS samples.](https://qlogix.ai/products/electronic-lab-notebook) - Environmental and water monitoringContamination control: viable and non-viable monitoring by cleanroom grade and water testing by point of use, judged against Annex 1 and pharmacopoeial limits, excursions investigated, results trended, each month reviewed and approved. - [Calibration management Every instrument, tool and gauge: an approved requirement, calibrations with readings judged and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [MES Electronic batch records from your own master batch record: every step executed and verified with credential signatures, entries out of limits flagged, QC samples in the LIMS, orders and consumption with the ERP, release signed by QA.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERP Materials, inventory lots with genealogy and QA's signed release, bills of materials as unit operations, work orders stage by stage to the finished lot, scheduling, budgets and the material plan; released lots lock dispensing in the MES.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooks Clean room, equipment and instrument logbooks kept as signed entries, with the periodic review the system checks first.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM Library A characterized, versioned particulate defect library for fill-finish: inspector qualification kits and AVI challenge sets that stay current.](https://qlogix.ai/products/pm-library) - [Stability Study Management ICH stability studies on the LIMS: barcode pulls at the chamber, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [CMMS Assets with preventive plans and work orders verified for return to service; reliability, lifecycle and predicted failures from the history.](https://qlogix.ai/products/cmms) ### Validation and projects - [Validation Plans, URS, risk assessments, protocols and reports from Part 11-ready templates, with GAMP 5 and CSA scaling the effort.](https://qlogix.ai/products/computer-system-validation) - [Operational excellence A3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment, computed by the system and coached by the AI; your own CI applications; OPEX projects in their own series.](https://qlogix.ai/products/operational-excellence) - [Project management Programs and projects on audited records: work breakdown, resources, risks, baseline, critical path and earned value.](https://qlogix.ai/products/project-management) ### Intelligence and automation - [Q Intelligence Ask across every record, connect your systems, draft the next record for a person to review.](https://qlogix.ai/products/ai-for-pharma-quality) - [Q Agent and agents Goals worked across every module, from the LIMS and ERP to validation and submissions, inside guardrails: every action waits for a person's approval, and every record drafted goes to its author for review.](https://qlogix.ai/products/ai-for-pharma-quality) - Q Dashboard and chartsThe quality system at a glance, and charts generated from your data, files or a live feed. - Regulatory libraryFDA, ICH and EU GMP with page-level citations, inside every answer. - AnalyticsTrends, capability and the metrics that show whether the system is in control. - Security and Part 11Hash-chained audit trails, electronic signatures with password re-entry, access levels and a security log on every record. - AI final reviewBefore any record is routed for approval, the AI reads it as the reviewers will and lists what to fix. Licensed per module and per user, ordered by purchase order; no card, no self-serve plan. [Ask for a quote](https://qlogix.ai/contact) ## What we implement Five practice areas, delivered separately or as one program. 01 ### QMS digitization Configured on the platform that fits your size and product type. - Deviations - CAPA - Change control - Complaints - Document control - Training - Supplier quality - APR/PQR [How QMS digitization works](https://qlogix.ai/services/qms-digitization) 02 ### AI-assisted investigations The AI drafts. Your investigators and QA decide. - IS / IS NOT problem statements - Fishbone - 5-Whys - CAPA effectiveness checks [How AI-assisted investigations works](https://qlogix.ai/services/ai-assisted-investigations) 03 ### Connected quality data Your QMS linked to plant and lab systems for trending, recurrence detection and earlier signals. - MES - LIMS - ERP - Batch records [How connected quality data works](https://qlogix.ai/services/connected-quality-data) 04 ### Validation of AI-enabled systems Risk-based validation following GAMP 5 (Second Edition) and computer software assurance principles, documented for inspection. - Intended use - Risk assessment - Test evidence - CSA [How validation of AI-enabled systems works](https://qlogix.ai/services/validation-of-ai-enabled-systems) 05 ### Agentic workflow automation Orchestrated agents that take on routine quality work across your systems: gathering, checking, drafting and chasing. Each one works inside defined guardrails and hands decisions to your people. - Orchestrator and specialist agents - Least-privilege access - Draft-only writes - Human approval gates - Audit trail [How agentic workflow automation works](https://qlogix.ai/services/agentic-workflow-automation) ## Built for sites without a large quality IT team ### Emerging and mid-size biotech and pharma Moving from paper or hybrid systems to a validated, connected QMS ahead of clinical or commercial scale. ### CDMOs and CMOs Managing multiple client quality agreements, audits and tech transfers on one system. ### 503B outsourcing facilities Meeting cGMP expectations with lean teams and high batch volume. ## How an engagement runs - 01 ### Assess A QMS readiness assessment: processes, systems, data flows and backlog. We agree baseline metrics before anything changes. - 02 ### Design Target architecture, platform selection, and a defined intended use and risk level for every AI function. - 03 ### Implement and validate Configuration, integration, risk-based validation, SOP updates and training. - 04 ### Monitor Performance monitoring, change control for model updates, and periodic review against the baseline. Most clients start with the assessment. [Book one](https://qlogix.ai/contact) ## AI in a decision-support role, by design Every workflow we build keeps GxP decisions with qualified people and leaves a record an inspector can follow. - Human review and approval for every GxP decision - Audit trails and traceability for AI-generated content - Data integrity to ALCOA+ expectations - Change control for model and prompt updates ### Frameworks we work to - 21 CFR Part 11 - 21 CFR Parts 210/211 - 21 CFR Part 820 (QMSR) - EU GMP Annex 11 - EU GMP Annex 22 (draft) - ICH Q9(R1) - ICH Q10 - GAMP 5 (2nd ed.) - FDA CSA guidance Annex 22 is a consultation draft at the time of writing. We design to its direction and update as final text is published. ## What we baseline and track We agree the measures in the assessment and report against them. - Cycle timeDeviation open to close 45days, open to closeBaselineTracked - Backlog ageOpen CAPAs and change controls 24records open over 90 daysBaselineTracked - RecurrenceRepeat deviations by root cause 8repeat deviations in the periodBaselineTracked - Right first timeBatch records and investigations 70% right first timeBaselineTracked Hover or tap a measure to see it move from baseline to tracked. The figures are illustrative; yours are measured in the assessment. ## Frequently asked questions ### About QLogix ### What does QLogix do? QLogix provides AI transformation for life sciences quality and operations in two ways: an AI applications platform for GxP operations (eQMS, root cause investigation, CAPA, complaints, change and document control, risk, regulatory submissions, LIMS, ELN, stability, MES, ERP, logbooks, calibration, CMMS, validation, project management and operational excellence), and implementation services that select, configure, validate and connect quality systems. AI drafts, trends and flags; qualified people review, decide and sign. ### Who is QLogix for? QLogix works with emerging and mid-size biotech and pharma companies moving from paper or hybrid systems to validated, connected ones; CDMOs and CMOs running several client quality agreements on one system; 503B outsourcing facilities meeting cGMP expectations with lean teams and high batch volume; and medical device and combination product manufacturers. ### What is an AI QMS? An AI QMS is a quality management system in which AI drafts records, investigations and assessments from the evidence on the record, quotes its sources and routes every decision to a qualified person, who approves with an electronic signature. In the QLogix Agentic QMS every AI contribution is an entry in the same hash-chained audit trail as the people's. ### The platform ### Which products does the QLogix platform include? Twenty-two modules in five groups. Quality events: root cause investigation, CAPA management, complaint handling and batch record review. Documents, change and risk: change management, document control, risk assessment (FMEA) and regulatory submissions. Laboratory, manufacturing and equipment: LIMS, electronic lab notebook, stability study management, MES with electronic batch records, ERP, GMP logbooks, the PM Library, calibration management and CMMS. Validation, projects and improvement: computer system validation, project management and operational excellence. AI for quality: Q Intelligence, Q Agent and agents. The Agentic QMS ties them together. ### Are the modules connected? Yes. Every QLogix module keeps its records on one audited record store, so a deviation links to the batch in the MES, the result in the LIMS, the work order in the CMMS and the CAPA that closes it, in both directions. Each record has a SHA-256 hash-chained audit trail, review and approval with electronic signatures, attachments and links, and each organization sees only its own data. ### Can we buy a single module? Yes. Several QLogix modules are sold on their own: for example root cause investigation for organizations that keep deviations and CAPAs in another QMS, or the CMMS, ELN, calibration management, project management, ERP or operational excellence. They connect to the rest of the platform when you add more. ### Does QLogix connect to our other systems? Yes. QLogix connects to ERP, LIMS, MES, BI tools, data warehouses, integration platforms such as Power Automate, Zapier, Workato or MuleSoft, and AI assistants, through a REST API and an MCP server. Keys are issued per client workspace on request and read that workspace's data only. ### Quality events ### How does root cause investigation work? QLogix Root cause investigation takes a deviation, OOS or OOT result, complaint or audit finding from the first write-up to a tested root cause and a corrective action plan. The AI drafts the problem statement with IS / IS NOT, the 5 Whys, the fishbone and a CAPA proposal from the write-up and documents; the investigator searches related records, builds the sequence of events and the change analysis, tests each cause as a hypothesis, scores the risk (ICH Q9) and the cost of poor quality, and decides. It works on its own or with the Agentic QMS. ### How is CAPA effectiveness shown? In QLogix CAPA management each corrective and preventive action has an owner, a due date and, when done, its evidence, and is rated on the hierarchy of controls so retraining alone is flagged. Effectiveness is measured from data: the counts before and after the actions give a one-sided exact test and an upper confidence bound, with the residual risk and a business case for the CAPA alongside. ### What do complaint handling and batch record review do? QLogix complaint handling turns a complaint as received into a structured record with quoted evidence, a route decided by a person and, where needed, a root cause investigation. QLogix batch record review reviews an executed batch record by exception, electronic or scanned paper, and lists only the entries that miss their instruction, limit or good documentation practice, for QA to disposition. ### Documents, change and risk ### How do change control and document control work? QLogix change management takes a change to a document, piece of equipment, area or utility from request to closure: the AI assesses its impact, risk and the validation it needs as soon as it is raised, and the functions affected sign off. QLogix document control manages SOPs, work instructions, specifications and master batch records with their versions and review periods; the AI reviews each draft, and approved versions become effective with read-and-understood training. ### How are risk assessments and regulatory submissions handled? The QLogix risk assessment module runs each design, process, equipment or CSV FMEA as one signed record, from the AI's first draft of the register to the acceptance of the residual risk, and investigations feed new failure modes back into it. The regulatory submissions module screens a change for regulatory impact and drafts submission text in which every claim quotes its source; regulatory affairs decides what is filed. ### Laboratory, manufacturing and equipment ### What do the LIMS, ELN and stability modules cover? The QLogix LIMS records in-process, release, stability and raw material results against approved specifications, calculates each result from the determinations and judges it against the limits, on instruments in calibration and standards in date. The electronic lab notebook is the analyst's contemporaneous record of runs, preparations and method work. Stability study management builds protocols on ICH Q1A(R2) conditions, schedules the pulls and trends the results. ### What do the MES and ERP do? The QLogix MES runs electronic batch records from your own master batch record, drafted by the AI from your existing record and approved by QA, for therapeutic products, medical devices and combination products. The QLogix ERP runs materials and products from the item master and purchasing through receipt, QC release, lot genealogy and inventory, on the same records as quality, the LIMS and the MES. ### What about equipment, logbooks and visual inspection? QLogix calibration management keeps one record per instrument or standard with its approved requirement and each calibration's result. The QLogix CMMS manages assets with their criticality, GxP impact and preventive maintenance, and work orders with labour, parts and downtime. GMP logbooks record cleaning, equipment use and instrument entries contemporaneously with second-person verification. The PM Library characterizes the particles and defects of parenteral fill-finish for visual inspection. ### Validation, projects and improvement ### How does computer system validation work in QLogix? The QLogix validation module runs the validation lifecycle on signed records for computerized systems, equipment, facilities and utilities, processes and cleaning: an AI-drafted validation plan, risk-based requirements and test scripts, executed tests with deviations, a traceability matrix and the summary report, following GAMP 5 (Second Edition) and computer software assurance. ### What do project management and operational excellence add? QLogix project management runs programs and projects with an AI-drafted plan, resources and load, risks, the critical path and earned value. QLogix operational excellence runs continuous improvement tools as applications: A3, kaizen, value stream maps, 5S, standard work, daily management, KPI bowlers and policy deployment, with computed results and action plans. A CAPA action can start the improvement tool that fits it, linked back. ### AI for quality ### What are Q Intelligence and Q Agent? QLogix Q Intelligence is an AI quality assistant that answers questions about any record, document, audit trail or regulation, shows every source it consulted and checks its citations; every answer is logged for validation. QLogix Q Agent carries out multi-step work across modules, and every action it proposes waits for a person's approval. ### Does the AI make quality decisions? No. In QLogix the AI drafts, summarizes, trends and flags. Review, approval, root cause conclusions, risk acceptance and batch disposition stay with qualified people, who sign electronically, and each AI contribution is recorded in the audit trail with the model and the sources it used. ### Services ### Which services does QLogix offer? Five services: QMS digitization (selecting, configuring, validating and migrating to an electronic QMS); AI-assisted investigations (problem statements, root cause candidates and CAPA plans drafted for investigators to approve); connected quality data (linking the QMS to MES, LIMS, ERP and batch records for trending); validation of AI-enabled systems (risk-based, to GAMP 5 and CSA); and agentic workflow automation (AI agents that gather, check, draft and chase within guardrails, with human approval). ### Do we have to replace our QMS? Usually not. QLogix is platform-agnostic: it works with your current QMS and its integrations with MES, LIMS and ERP, and recommends a change of platform only when the assessment shows the current one cannot meet the need. Modules such as root cause investigation can sit beside an existing QMS. ### How are AI functions validated? QLogix validates AI functions with a risk-based approach: defined intended use, a risk assessment, documented testing against acceptance criteria, monitoring in use, and change control for any model or prompt update. Q Intelligence comes with an intended use, a risk assessment, an automated controls check and an append-only answer log as validation evidence. ### Compliance ### Which regulations and frameworks does QLogix work to? QLogix works to 21 CFR Part 11, 21 CFR Parts 210 and 211, 21 CFR Part 820 (QMSR), EU GMP Annex 11, the draft EU GMP Annex 22 on AI, ICH Q9(R1), ICH Q10, ICH Q1A(R2) for stability, GAMP 5 (Second Edition) and FDA's computer software assurance guidance, with ALCOA+ data integrity throughout. ### How does QLogix meet Part 11 and data integrity expectations? Every QLogix record has an append-only audit trail chained with SHA-256 for tamper evidence, electronic signatures with re-authentication and the meaning of the signature, role-based access with review and approval separated by default, and attributable, time-stamped entries. Records are corrected with a reason, never deleted, and exports of the controlled copy are logged. ### Working with QLogix ### How does an engagement start? Most QLogix clients start with a QMS readiness assessment covering processes, systems, data flows and backlog. Baseline metrics such as deviation cycle time, backlog age, recurrence and right first time are agreed before anything changes, so the improvement can be measured. ### Where does QLogix work? QLogix is based in Newport Beach, California, and works on site across Southern California, with remote delivery elsewhere in the United States. ## AI transformation for life sciences quality, topic by topic QLogix delivers AI transformation solutions for life sciences: AI QMS software and QMS implementation for pharma, biotech, CDMOs and 503B facilities. One validated platform covers eQMS, deviation and CAPA management, change and document control, complaints, LIMS, the electronic lab notebook, electronic batch records, GMP logbooks, CMMS, calibration and computer system validation, built to 21 CFR Part 11 and EU GMP Annex 11. ### Services - [QMS digitization](https://qlogix.ai/services/qms-digitization) QMS digitization · QMS implementation · paper to eQMS · eQMS selection · QMS configuration · QMS validation · QMS migration - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) AI-assisted investigations · AI root cause analysis · deviation investigation · CAPA effectiveness · 5 Whys and fishbone analysis · AI for deviations - [Connected quality data](https://qlogix.ai/services/connected-quality-data) connected quality data · QMS integration · LIMS, MES and ERP integration · quality data integration · quality metrics · data integrity - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) validation of AI-enabled systems · AI validation · AI model validation for pharma · GAMP 5 for AI · computer software assurance · EU GMP Annex 22 readiness - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) agentic workflow automation · AI agents for quality · agentic AI in GxP · quality workflow automation · human-in-the-loop AI ### Quality events - [Root cause investigation](https://qlogix.ai/products/root-cause-investigation) root cause analysis software · root cause investigation software · deviation investigation software · OOS investigation software · AI root cause analysis · 5 Whys and fishbone analysis · GMP deviation investigation - [CAPA management](https://qlogix.ai/products/deviation-capa-management) CAPA software · CAPA management software · CAPA effectiveness check · corrective and preventive action system · GMP CAPA - [Complaint handling](https://qlogix.ai/products/complaint-handling) complaint handling software · complaint management system · pharmaceutical complaint handling · medical device complaint handling · 21 CFR 211.198 complaint files · 21 CFR 820.198 complaint handling · adverse event triage · MDR and field alert reportability assessment - [Batch record review](https://qlogix.ai/products/batch-record-review) batch record review software · review by exception · batch record review by exception · executed batch record review · batch release review · QA batch disposition · electronic batch record review · GMP batch record review What it looks like ### Signal detection: a drift caught before the limit Fill weight over 16 lots. The run rules flag the shift at lot 10 ; the specification is still 7 lots away at this rate. Illustrative figures from the simulations on this page, not client results. ### Documents, change and risk - [Change management](https://qlogix.ai/products/change-control) change control software · GMP change management software · GMP change control · change management for pharma · equipment change control GMP · facility change control GMP · document change control · change impact assessment · regulatory impact of changes · change control effectiveness check · ICH Q9 change risk assessment · pharmaceutical change control system - [Document control](https://qlogix.ai/products/document-control) document control software · GMP document management system · SOP management software · controlled documents · document periodic review · read and understood training · pharmaceutical document control · electronic document management for life sciences - [Risk assessment](https://qlogix.ai/products/quality-risk-management-fmea) quality risk management software · FMEA software · ICH Q9(R1) · pFMEA · dFMEA · equipment FMEA · CSV FMEA · risk priority number · risk assessment for pharma · mitigation tracking - [Regulatory submissions](https://qlogix.ai/products/regulatory-submissions) regulatory submissions software · post-approval changes · regulatory impact assessment · CMC changes · eCTD · NDA dossier drafting · BLA dossier drafting · CBE-30 and prior approval supplement · EU variations ### Laboratory, manufacturing and equipment - [LIMS](https://qlogix.ai/products/lims) LIMS for pharma · LIMS software for QC laboratories · pharmaceutical LIMS · OOS management software · out-of-specification investigation · certificate of analysis software · reference standard management · instrument calibration lockout · GMP laboratory data integrity - [Electronic lab notebook](https://qlogix.ai/products/electronic-lab-notebook) electronic lab notebook · ELN for QC laboratories · GMP electronic lab notebook · analytical assay documentation · ELN LIMS integration · laboratory data integrity · QC lab software - [Stability Study Management](https://qlogix.ai/products/stability-study-management) stability study management software · stability testing software · ICH Q1E shelf life calculation · shelf life estimation software · stability pull schedule · stability sample management · stability LIMS · stability chamber management · out of trend stability · ICH Q1A stability protocol · stability data trending · retest period estimation - [MES (electronic batch records)](https://qlogix.ai/products/mes-electronic-batch-records) MES for pharma · electronic batch records · EBR software · eBR system · master batch record · review by exception · manufacturing execution system · MES ERP integration · batch release - [ERP](https://qlogix.ai/products/erp-materials-inventory) GMP ERP · pharma ERP software · ERP for pharmaceutical manufacturing · inventory management for GMP · lot genealogy · QC on receipt · bill of materials software · work order management pharma · material requirements planning GMP · production scheduling pharma · MES ERP integration · LIMS ERP integration - [GMP logbooks](https://qlogix.ai/products/gmp-electronic-logbooks) electronic logbooks for GMP · GMP electronic logbook · equipment logbook software · cleanroom cleaning log · instrument logbook · logbook periodic review · ALCOA+ data integrity · eLogbook - [PM Library](https://qlogix.ai/products/pm-library) particulate matter library · visual inspection qualification · AVI challenge set · defect library for fill-finish · inspector requalification · visible particulates in injections · USP 790 and 1790 · Annex 1 visual inspection · FTIR particle identification - [Calibration management](https://qlogix.ai/products/calibration-management) calibration management software · GMP calibration · instrument calibration tracking · as-found and as-left results · out-of-tolerance impact assessment · calibration scheduling · metrology software - [CMMS](https://qlogix.ai/products/cmms) CMMS for GMP · GMP maintenance software · computerized maintenance management system · preventive maintenance software · work order management · asset management for pharma · predictive maintenance · MTBF and MTTR · equipment reliability What it looks like ### Deviation cycle time, by stage Days per stage when the evidence is retyped from MES, LIMS and ERP, against the same stages with the systems connected and the AI drafting from one record. Review and approval stay with people, so they shrink the least. Gather the evidence9 → 1 d Draft the investigation11 → 2 d Review and approve14 → 9 d CAPA plan agreed7 → 4 d Open to closed41 d → 16 d Illustrative figures from the simulations on this page, not client results. ### Validation, projects and improvement - [Validation](https://qlogix.ai/products/computer-system-validation) computer system validation software · CSV and CSA · computer software assurance · GAMP 5 validation · paperless validation · validation lifecycle management · IQ OQ PQ protocols · electronic test execution · traceability matrix - [Project management](https://qlogix.ai/products/project-management) GxP project management · validation project management · critical path scheduling · earned value management · validation program tracking · project resource planning - [Operational excellence](https://qlogix.ai/products/operational-excellence) operational excellence software · continuous improvement software · lean management software · A3 problem solving software · kaizen software · value stream mapping software · 5S audit software · standard work software · daily management board · KPI bowler · hoshin kanri software · policy deployment X-matrix · OPEX project management ### AI for quality - [Agentic QMS](https://qlogix.ai/products/agentic-qms) eQMS software · QMS software for life sciences · AI QMS · agentic QMS · electronic quality management system · pharma QMS software · quality management software for biotech · medical device QMS software · cloud QMS - [Q Intelligence, Q Agent and agents](https://qlogix.ai/products/ai-for-pharma-quality) AI for pharmaceutical quality · AI in GxP · generative AI for QMS · AI agents for quality assurance · AI quality assistant · AI with human approval · EU GMP Annex 22 · FDA AI credibility framework What it looks like ### CAPAs closed, not carried Open CAPAs over twelve months once every action has an owner, a due date and a reminder, and each effectiveness check is scheduled from the record. Open CAPAs 38 → 14 Overdue actions 9 → 1 Effectiveness checks on time 61% → 96% Illustrative figures from the simulations on this page, not client results. ## From the blog: AI and quality in life sciences - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.Read the article](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.Read the article](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.Read the article](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) [All articles](https://qlogix.ai/blog) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # QMS and AI quality services for life sciences | QLogix Source: https://qlogix.ai/services # QMS and AI quality services for life sciences QLogix works in five practice areas, delivered separately or as one program. Each starts from the quality system you already have and keeps every GxP decision with qualified people. - [01QMS digitizationQMS digitization is the move from paper or hybrid quality processes to a validated electronic quality management system. QLogix selects the platform that fits your size and product type, configures it to your SOPs, validates it and migrates the records you need to keep.Read about QMS digitization](https://qlogix.ai/services/qms-digitization) - [02AI-assisted investigationsAn AI-assisted investigation uses AI to prepare the groundwork of a deviation investigation: the problem statement, similar past events, root cause candidates and a draft CAPA plan. The investigator and QA review, edit and approve every part. The AI drafts; it does not conclude.Read about AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [03Connected quality dataConnected quality data means linking the QMS to the plant and lab systems around it (MES, LIMS, ERP and electronic batch records) so that quality events, batch data and lab results can be trended together. It turns a QMS that records everything into one that also shows patterns before they become recurrences.Read about connected quality data](https://qlogix.ai/services/connected-quality-data) - [04Validation of AI-enabled systemsValidating an AI-enabled system means showing, with documented evidence, that each AI function is fit for its intended use and stays that way. QLogix follows GAMP 5 (Second Edition) and computer software assurance principles: effort is scaled to the risk of the function, and the evidence is written for an inspector to follow.Read about validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) - [05Agentic workflow automationAgentic workflow automation uses orchestrated AI agents to carry out routine quality work across your systems: gathering, checking, drafting and chasing. Each agent works inside defined guardrails, can draft but not approve, and hands decisions to your people.Read about agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # QMS digitization for biotech, pharma, CDMOs and 503B | QLogix Source: https://qlogix.ai/services/qms-digitization # QMS digitization for life sciences QMS digitization is the move from paper or hybrid quality processes to a validated electronic quality management system. QLogix selects the platform that fits your size and product type, configures it to your SOPs, validates it and migrates the records you need to keep. [Request a meeting](https://qlogix.ai/contact)[All services](https://qlogix.ai/services) ## What we configure The core quality processes, set up as connected workflows rather than separate forms, so a deviation can open a CAPA, a CAPA can open a change control, and each one links back to the records behind it. - Deviations and investigations - CAPA, with effectiveness checks - Change control - Complaints - Document control and training - Supplier quality - Annual product review (APR/PQR) data ## How the work runs We start with a readiness assessment of your processes, systems, data flows and backlog, and agree baseline metrics before anything changes. Platform selection follows from that assessment. We are platform-agnostic, and where your current QMS can meet the need we configure and connect it instead of replacing it. Configuration, integration and risk-based validation run together with SOP updates and training, so the system goes live with procedures that describe how it is actually used. ## What makes it hold up in an inspection - Electronic records and signatures set up to meet 21 CFR Part 11 and EU GMP Annex 11 - Audit trails that show who changed what, when and why - Data integrity to ALCOA+ expectations - Validation evidence scaled to risk, following GAMP 5 (Second Edition) ## The products that deliver it QMS digitization runs on these modules of the QLogix platform. - [Root cause investigationDeviation, OOS and complaint investigations with AI-drafted problem statements, hypotheses tested and a robust root cause; on its own beside your existing QMS.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementCAPAs opened from investigations, complaints, audits and changes, with an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Change managementChanges to documents, equipment, facilities and utilities under GMP: what changes and how, AI impact and risk assessment, the GMP requirements each item must meet, tasks, approvals to implement, effectiveness and closure.](https://qlogix.ai/products/change-control) - [Document managementControlled documents with versions, AI review, approvals, training, effective dates and periodic review; every QMS document in one index.](https://qlogix.ai/products/document-control) - [Complaint handlingTriage with the adverse-event check, evidence quoted from the complaint, and the decision signed.](https://qlogix.ai/products/complaint-handling) - [Risk assessmentDesign, process, equipment and CSV FMEA with RPN thresholds, mitigation actions and reassessment.](https://qlogix.ai/products/quality-risk-management-fmea) - [Batch record reviewExceptions found, assessed and dispositioned, with the recommendation signed by QA.](https://qlogix.ai/products/batch-record-review) ## Your qualified people decide - Platform choice and configuration decisions are reviewed and approved by your quality unit. - Workflow approvals, dispositions and signatures stay with the named, qualified people in your procedures. ## Frameworks this work is built to - 21 CFR Part 11 - 21 CFR Parts 210/211 - 21 CFR Part 820 (QMSR) - EU GMP Annex 11 - ICH Q10 - GAMP 5 (2nd ed.) ## Included - Deviations - CAPA - Change control - Complaints - Document control - Training - Supplier quality - APR/PQR ## Questions about QMS digitization ### Do we have to replace our current QMS? Usually not. We start from what you have and recommend a change of platform only when the assessment shows the current one cannot meet the need. ### Which QMS platforms do you implement? We are platform-agnostic and work across the major QMS platforms and their integrations with MES, LIMS and ERP. The platform is chosen in the assessment, based on your size and product type. ### What happens to our paper and legacy records? The assessment identifies which records have to be migrated, which can stay in their original form under your retention procedure, and how the migration will be verified. ### Who is QMS digitization for? Emerging and mid-size biotech and pharma companies moving to a validated QMS ahead of clinical or commercial scale, CDMOs and CMOs running several client quality agreements on one system, and 503B outsourcing facilities with lean teams and high batch volume. ## Related reading - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Related services - [AI-assisted investigationsThe AI drafts. Your investigators and QA decide.](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality dataYour QMS linked to plant and lab systems for trending, recurrence detection and earlier signals.](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systemsRisk-based validation following GAMP 5 (Second Edition) and computer software assurance principles, documented for inspection.](https://qlogix.ai/services/validation-of-ai-enabled-systems) - [Agentic workflow automationOrchestrated agents that take on routine quality work across your systems: gathering, checking, drafting and chasing. Each one works inside defined guardrails and hands decisions to your people.](https://qlogix.ai/services/agentic-workflow-automation) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI-assisted deviation investigations and CAPA | QLogix Source: https://qlogix.ai/services/ai-assisted-investigations # AI-assisted deviation investigations and CAPA An AI-assisted investigation uses AI to prepare the groundwork of a deviation investigation: the problem statement, similar past events, root cause candidates and a draft CAPA plan. The investigator and QA review, edit and approve every part. The AI drafts; it does not conclude. [Request a meeting](https://qlogix.ai/contact)[All services](https://qlogix.ai/services) ## What the AI drafts - The problem statement in IS / IS NOT form, from the initial report - Similar past events across batches, equipment and sites - Root cause candidates by fishbone category, each with the evidence it came from - 5-Whys chains for the candidates your investigator keeps - A draft CAPA plan with owners, timing and effectiveness checks ## Why it shortens cycle time Most of the elapsed time in an investigation goes to gathering: finding the batch record pages, the earlier deviations, the lab results and the procedure in force. When that arrives as a sourced draft, the investigator starts by checking and deciding instead of starting from a blank page, and the quality of the write-up depends less on who was available. We baseline deviation cycle time, backlog age, recurrence and right first time in the assessment, and report against the same measures afterwards. ## How it stays inspection-ready 21 CFR 211.192 requires unexplained discrepancies to be thoroughly investigated, with a written record of the conclusions and follow-up. The AI's part in that record is visible, not hidden. - Every AI draft is labelled as AI-generated and quotes the document it relied on - Edits, acceptances and rejections are recorded in the audit trail with a reason - The record is signed by the people accountable for it - Models and prompts are under change control ## The products that deliver it AI-assisted investigations runs on these modules of the QLogix platform. - [Root cause investigationDeviation, OOS and complaint investigations with AI-drafted problem statements, hypotheses tested and a robust root cause; on its own beside your existing QMS.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementCAPAs opened from investigations, complaints, audits and changes, with an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingTriage with the adverse-event check, evidence quoted from the complaint, and the decision signed.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewExceptions found, assessed and dispositioned, with the recommendation signed by QA.](https://qlogix.ai/products/batch-record-review) - [Q IntelligenceAsk across every record, connect your systems, draft the next record for a person to review.](https://qlogix.ai/products/ai-for-pharma-quality) ## Your qualified people decide - Classification, product impact and reportability are decided by qualified people. - Root cause is concluded by the investigator and approved by QA. - CAPA actions are accepted or rejected one by one, and the record is signed by a person. ## Frameworks this work is built to - 21 CFR 211.192 - 21 CFR Part 11 - ICH Q9(R1) - ICH Q10 - EU GMP Annex 11 ## Included - IS / IS NOT problem statements - Fishbone - 5-Whys - CAPA effectiveness checks ## Questions about AI-assisted investigations ### Does the AI decide the root cause? No. It proposes candidates with the evidence for each. The investigator rules candidates in or out, concludes the root cause, and QA approves it. ### Is AI allowed in GMP investigations? The regulations require a thorough, documented investigation and accountable review by the quality unit. They do not specify the tools used to prepare it. The controls that matter are a defined intended use, validation scaled to risk, human review and approval of every GxP decision, and an audit trail of what the AI contributed. ### What data does the AI read? Only the sources you connect and allow: typically the deviation report, batch records, lab results, procedures and earlier quality records. Access is least-privilege and every source read is recorded. ### How do you know the AI draft is right? Each statement in a draft is tied to a quoted source, so a reviewer can check it. The function is tested against acceptance criteria during validation and monitored afterwards, and a draft is never the final record until a qualified person has reviewed and approved it. ## Related reading - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 14 October 2026Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Related services - [QMS digitizationConfigured on the platform that fits your size and product type.](https://qlogix.ai/services/qms-digitization) - [Connected quality dataYour QMS linked to plant and lab systems for trending, recurrence detection and earlier signals.](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systemsRisk-based validation following GAMP 5 (Second Edition) and computer software assurance principles, documented for inspection.](https://qlogix.ai/services/validation-of-ai-enabled-systems) - [Agentic workflow automationOrchestrated agents that take on routine quality work across your systems: gathering, checking, drafting and chasing. Each one works inside defined guardrails and hands decisions to your people.](https://qlogix.ai/services/agentic-workflow-automation) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # QMS integration with MES, LIMS and ERP | QLogix Source: https://qlogix.ai/services/connected-quality-data # Connected quality data: QMS integration with MES, LIMS and ERP Connected quality data means linking the QMS to the plant and lab systems around it (MES, LIMS, ERP and electronic batch records) so that quality events, batch data and lab results can be trended together. It turns a QMS that records everything into one that also shows patterns before they become recurrences. [Request a meeting](https://qlogix.ai/contact)[All services](https://qlogix.ai/services) ## The problem it solves Most sites run a capable QMS next to MES, LIMS and ERP systems that never talk to it. Trends sit across batch records, lab results and complaints, and nobody sees the pattern until it becomes a repeat deviation. ## What we connect - MES and electronic batch records: process parameters, exceptions and yields - LIMS: results, out-of-specification and out-of-trend events - ERP: materials, lots, suppliers and genealogy - QMS: deviations, CAPAs, change controls and complaints ## What you can do once it is connected - Detect sustained shifts and trends in batch and lab results before a limit is crossed - Find repeat deviations by root cause across products, lines and sites - Review batch records by exception instead of page by page - Assemble annual product review (APR/PQR) data from the systems of record ## How the interfaces are controlled Each interface has a defined purpose, an owner and documented testing. Data keeps its source, timestamp and attribution as it moves, so a figure in a trend can be traced back to the record it came from. Changes to an interface go through change control. ## The products that deliver it Connected quality data runs on these modules of the QLogix platform. - [LIMSSpecifications, samples and results judged by the system, OOS handled under investigation, instruments and standards locked out when out of date.](https://qlogix.ai/products/lims) - [MESElectronic batch records from your own master batch record: every step executed and verified with credential signatures, entries out of limits flagged, QC samples in the LIMS, orders and consumption with the ERP, release signed by QA.](https://qlogix.ai/products/mes-electronic-batch-records) - [Electronic lab notebookEach analytical run recorded as it happens: the effective method, equipment checked across the LIMS, calibration and maintenance, standards and reagents in date, observations never edited, results posted to the LIMS samples.](https://qlogix.ai/products/electronic-lab-notebook) - [Signal detectionControl charts with run rules on any parameter; each signal interpreted by the AI and dispositioned by QA.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [Batch record reviewExceptions found, assessed and dispositioned, with the recommendation signed by QA.](https://qlogix.ai/products/batch-record-review) - [Stability Study ManagementICH stability studies on the LIMS: barcode pulls at the chamber, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Q IntelligenceAsk across every record, connect your systems, draft the next record for a person to review.](https://qlogix.ai/products/ai-for-pharma-quality) ## Your qualified people decide - A flagged signal is a prompt to look, not a conclusion. QA decides whether it is a deviation, a trend to watch or nothing. - Batch disposition stays with the quality unit. ## Frameworks this work is built to - 21 CFR Part 11 - 21 CFR 211.68 - 21 CFR 211.180(e) - EU GMP Annex 11 - GAMP 5 (2nd ed.) ## Included - MES - LIMS - ERP - Batch records ## Questions about connected quality data ### Do we need a data lake or a new platform first? Not necessarily. The assessment maps the data flows you already have and the questions you need answered. Often a small number of well-controlled interfaces between the QMS and the systems of record is enough to start. ### Does integrating systems mean revalidating all of them? No. Validation is scaled to the risk of each interface and its intended use. The systems of record keep their validated state; the interface and the use of its data are what get assessed, tested and documented. ### What is signal detection in a quality context? It is the use of control-chart rules and trending on batch and lab data to flag sustained shifts, trends and excursions early. The system explains why a point was flagged; QA decides what to do about it. ## Related reading - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [AI in quality · 14 October 2026Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Related services - [QMS digitizationConfigured on the platform that fits your size and product type.](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigationsThe AI drafts. Your investigators and QA decide.](https://qlogix.ai/services/ai-assisted-investigations) - [Validation of AI-enabled systemsRisk-based validation following GAMP 5 (Second Edition) and computer software assurance principles, documented for inspection.](https://qlogix.ai/services/validation-of-ai-enabled-systems) - [Agentic workflow automationOrchestrated agents that take on routine quality work across your systems: gathering, checking, drafting and chasing. Each one works inside defined guardrails and hands decisions to your people.](https://qlogix.ai/services/agentic-workflow-automation) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Validation of AI-enabled GxP systems (GAMP 5, CSA) | QLogix Source: https://qlogix.ai/services/validation-of-ai-enabled-systems # Validation of AI-enabled GxP systems Validating an AI-enabled system means showing, with documented evidence, that each AI function is fit for its intended use and stays that way. QLogix follows GAMP 5 (Second Edition) and computer software assurance principles: effort is scaled to the risk of the function, and the evidence is written for an inspector to follow. [Request a meeting](https://qlogix.ai/contact)[All services](https://qlogix.ai/services) ## What gets defined for every AI function - Intended use: what the function does, for whom, and what it must never do - Risk level: what happens if the output is wrong, and who would catch it - The human control: who reviews and approves the output before it becomes a record - Acceptance criteria: what good enough looks like, in measurable terms ## What the evidence looks like - A risk assessment that drives how much testing each function gets - Documented testing against the acceptance criteria, on representative cases - Checks that outputs are sourced, labelled as AI-generated and traceable - Audit trail and access control verification - A validation report that states what was tested, what was found and what is monitored ## Keeping the validated state An AI function changes when its model, its prompt or its data sources change. Each of those goes through change control with an impact assessment and regression testing in proportion to risk. Performance is monitored in use and reviewed periodically against the baseline. EU GMP Annex 22 on artificial intelligence is a consultation draft at the time of writing. We design to its direction and update as final text is published. ## The products that deliver it Validation of AI-enabled systems runs on these modules of the QLogix platform. - [ValidationPlans, URS, risk assessments, protocols and reports from Part 11-ready templates, with GAMP 5 and CSA scaling the effort.](https://qlogix.ai/products/computer-system-validation) - [Risk assessmentDesign, process, equipment and CSV FMEA with RPN thresholds, mitigation actions and reassessment.](https://qlogix.ai/products/quality-risk-management-fmea) - [Q IntelligenceAsk across every record, connect your systems, draft the next record for a person to review.](https://qlogix.ai/products/ai-for-pharma-quality) ## Your qualified people decide - Your quality unit approves the intended use, the risk assessment and the validation report. - A function is released for use by a person, not by passing its tests. ## Frameworks this work is built to - GAMP 5 (2nd ed.) - FDA CSA guidance - 21 CFR Part 11 - EU GMP Annex 11 - EU GMP Annex 22 (draft) - ICH Q9(R1) ## Included - Intended use - Risk assessment - Test evidence - CSA ## Questions about validation of AI-enabled systems ### How do you validate something that does not give the same answer twice? By validating the function in its context of use. A drafting function with mandatory human review is tested for whether its outputs are sourced, complete enough to be useful and correctly labelled, across representative cases and against acceptance criteria, instead of for identical output on every run. ### What is the difference between CSV and CSA? Computerized system validation (CSV) has often been practiced as scripted testing of everything. Computer software assurance (CSA) is FDA's risk-based approach: put the most rigorous assurance where a failure would affect product quality or patient safety, and use lighter, unscripted testing where it would not. ### What happens when the AI model is updated? The update is handled as a change: impact assessment, regression testing against the acceptance criteria in proportion to risk, approval, and a record of the model version in use for each output. ### Can you validate an AI function we already use? Yes. The starting point is the same: write down its intended use and risk, find out what evidence already exists, and close the gaps. ## Related reading - [Validation · 14 October 2026AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. The same holds for validation: plans, protocols and reports drafted from templates and requirements, with assurance effort set by risk under GAMP 5 and FDA's CSA guidance, and a person approving every version with an electronic signature.](https://qlogix.ai/blog/ai-document-management-validation-workflow) - [AI in quality · 12 October 2026FDA's seven-step credibility framework for AI, in plain termsFDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the decision, and how serious a wrong decision would be.](https://qlogix.ai/blog/fda-ai-credibility-framework) - [Manufacturing science · 12 October 2026AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing models. Together they show what a process model has to prove.](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) ## Related services - [QMS digitizationConfigured on the platform that fits your size and product type.](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigationsThe AI drafts. Your investigators and QA decide.](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality dataYour QMS linked to plant and lab systems for trending, recurrence detection and earlier signals.](https://qlogix.ai/services/connected-quality-data) - [Agentic workflow automationOrchestrated agents that take on routine quality work across your systems: gathering, checking, drafting and chasing. Each one works inside defined guardrails and hands decisions to your people.](https://qlogix.ai/services/agentic-workflow-automation) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Agentic AI workflow automation for GxP quality teams | QLogix Source: https://qlogix.ai/services/agentic-workflow-automation # Agentic workflow automation for quality teams Agentic workflow automation uses orchestrated AI agents to carry out routine quality work across your systems: gathering, checking, drafting and chasing. Each agent works inside defined guardrails, can draft but not approve, and hands decisions to your people. [Request a meeting](https://qlogix.ai/contact)[All services](https://qlogix.ai/services) ## How the architecture works An orchestrator agent plans the task, routes parts of it to specialist agents and assembles the result. The specialists each read the systems they are allowed to: an intake agent on the complaint inbox, an investigation agent on the QMS and batch records, a trending agent on LIMS and MES, a batch review agent on electronic batch records, and a document agent on the document system. ## Routine work agents take on - Preparing the morning quality huddle: what changed, what needs attention today - Chasing overdue CAPA actions, with a drafted reminder for each owner - Assembling the annual product review (APR/PQR) data pack - Triaging incoming complaints into a drafted record with quoted evidence - Ranking an open backlog by risk, with the reason for each flag ## The guardrails - Least-privilege access for each agent - Draft-only writes: an agent can propose a change, an email or a flag, and cannot make it final - Human approval gates before anything is sent or recorded - Every tool call in the audit trail, so an inspector can follow how an output was produced - Models and prompts under change control ## The products that deliver it Agentic workflow automation runs on these modules of the QLogix platform. - [Q Agent and agentsGoals worked across every module, from the LIMS and ERP to validation and submissions, inside guardrails: every action waits for a person's approval, and every record drafted goes to its author for review.](https://qlogix.ai/products/ai-for-pharma-quality) - [Q IntelligenceAsk across every record, connect your systems, draft the next record for a person to review.](https://qlogix.ai/products/ai-for-pharma-quality) - [Backlog triageOpen deviations, CAPAs and change controls ranked by risk, with QA's priorities on record.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [Change managementChanges to documents, equipment, facilities and utilities under GMP: what changes and how, AI impact and risk assessment, the GMP requirements each item must meet, tasks, approvals to implement, effectiveness and closure.](https://qlogix.ai/products/change-control) ## Your qualified people decide - Approvals, dispositions and reportability calls stay with qualified people. Agents bring them the work. - Each agent has a defined intended use that your quality unit has approved. ## Frameworks this work is built to - 21 CFR Part 11 - EU GMP Annex 11 - EU GMP Annex 22 (draft) - GAMP 5 (2nd ed.) - ICH Q9(R1) ## Included - Orchestrator and specialist agents - Least-privilege access - Draft-only writes - Human approval gates - Audit trail ## Questions about agentic workflow automation ### What is an AI agent in a quality system? A software component that uses an AI model to plan and carry out a bounded task with tools: reading records, searching documents, drafting text. In a quality system each agent has a defined intended use, only the access it needs, and no authority to approve. ### Can an agent change or approve a GxP record? No. Agents read and draft. Anything that would change a record, send a message or record a decision waits for a qualified person to review and approve it, and the approval is attributed to that person. ### How is this different from robotic process automation? Robotic process automation replays fixed steps in fixed screens. An agent can read unstructured records, decide which sources it needs and write a sourced draft. That flexibility is why it is bounded by access limits, draft-only writes and approval gates. ### Which task should we automate first? One that is routine, well understood and low risk if a draft is wrong: huddle preparation and overdue chasing are common starting points, because a person reads the result before acting on it. ## Related reading - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [Validation · 14 October 2026AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. The same holds for validation: plans, protocols and reports drafted from templates and requirements, with assurance effort set by risk under GAMP 5 and FDA's CSA guidance, and a person approving every version with an electronic signature.](https://qlogix.ai/blog/ai-document-management-validation-workflow) ## Related services - [QMS digitizationConfigured on the platform that fits your size and product type.](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigationsThe AI drafts. Your investigators and QA decide.](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality dataYour QMS linked to plant and lab systems for trending, recurrence detection and earlier signals.](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systemsRisk-based validation following GAMP 5 (Second Edition) and computer software assurance principles, documented for inspection.](https://qlogix.ai/services/validation-of-ai-enabled-systems) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI QMS software for life sciences: all products | QLogix Source: https://qlogix.ai/products # AI QMS software for life sciences, module by module QLogix is one eQMS and operations platform for pharma, biotech, CDMOs, medical devices and 503B facilities: deviations and CAPA, complaints, change control, document control, LIMS, electronic batch records, calibration, maintenance and validation on one set of audited records, with AI that drafts and flags while qualified people decide and sign. Each module is licensed on its own or with the others. [Agentic QMS: the platform that joins every moduleeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) ## Quality events Deviations, OOS results, complaints and batch records, investigated and closed with the evidence on the record. - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) ## Documents, change and risk The controlled documents, changes, risk assessments and regulatory filings that keep the validated state. - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) ## Laboratory, manufacturing and equipment The lab, the production floor and the equipment, on the same records as quality. - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) ## Validation, projects and improvement Computer system validation, continuous improvement and the projects that deliver them, on schedule and on the record. - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) ## AI for quality Answers, drafts and agents across every record, with a person approving every action. - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # eQMS software for life sciences: the Agentic QMS | QLogix Source: https://qlogix.ai/products/agentic-qms # Agentic QMS: eQMS software for pharma, biotech and medical devices The QLogix Agentic QMS is an electronic quality management system (eQMS) that keeps change control, CAPA and controlled documents on the same audited, signed records as every other QLogix module. Every create, change, decision, signature, export and AI draft is an entry in one append-only audit trail, chained with SHA-256 for the organization, so a deviation, the CAPA it opened, the change control that carries the action and the procedure that changed are one connected file. The AI assesses changes, reviews CAPA plans and reviews document drafts, quoting its sources; people decide, and only people sign. A work queue shows every open record across the platform, what each one is waiting for and who owns it. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Change control from proposal to closure: an AI impact assessment across nine areas with the ICH Q9 risk and quoted evidence, tasks owned by function, a signed approval to implement, an effectiveness check and a signed closure. - CAPA management: a CAPA opened from an approved investigation carries in the problem statement, root cause, classification and accepted actions; the AI reviews the plan for adequacy, gaps and extent of condition and proposes an effectiveness method, criterion and timeframe. - A rule inspectors expect: retraining cannot be the only corrective or preventive action, and a plan that relies on it alone cannot be approved. - Controlled documents with versions, numbered by type series (SOP, WI, POL, FRM, SPEC, PRT, RPT, MBR, DOC); the AI reviews each draft version, the findings are dispositioned, approval makes the version effective on its date, and trained people sign read and understood. - Stages derived from what the record contains, never set by hand: a change control moves from Proposed to Assessed, Approved, Implementing and Closed; a CAPA from Plan to Plan approved, Implementing, Effectiveness check and Closed. - A work queue of every open record across the platform (complaints, investigations, batch reviews, submissions, signals, change controls, CAPAs, documents and validation) with its stage, what it is waiting for, owner, due date and age, filtered by product or by what is overdue or due within 14 days. - An overview of the risk posture: changes awaiting a decision, emergency or expedited changes open, CAPAs past their target date or awaiting effectiveness, and documents due for periodic review. - Connected records linked by meaning (Originates from, Resulted in, Implements, Affects, Supersedes, Related to), shown from both sides, with suggestions of related records that share lots, equipment, documents, products or cause wording. - An AI final review before any record is routed: the AI reads the change control, CAPA or document as the reviewers will and lists what to fix, where and how; the author decides what to accept. ## How it fits the platform The Agentic QMS is the hub of the other modules rather than a separate system next to them. A complaint from Complaint handling or a deviation starts an investigation, and the approved investigation opens its CAPA prefilled; the CAPA raises the change control; the change control names the documents and training affected and the procedure is revised in document control; a change that may need a filing opens a regulatory impact screening in Regulatory submissions. Backlog triage ranks the open change controls and CAPAs by risk, Document management indexes every QMS document in one search, a modification in the CMMS goes through the same change control, and a validation project carries the reference of the change control behind it. Because there is one record store and one audit trail, nothing is copied between systems to be reconciled later. Q Intelligence answers questions across every record with citations, and Q Agent works a goal across the QMS, proposing each action for a person to approve. Every record page shows the audit trail integrity check, and the read-only REST API and MCP server let other systems and AI assistants read the same records under an API key. ## Where people decide The AI drafts: the change assessment, the CAPA plan review, the document review and the final review, each with exact quotes from the write-up or the attached documents. People choose every disposition, decision and route, with a reason when a value changes, and only people submit, review and approve, by electronic signature. The AI never signs, closes, releases a batch or decides reportability, and every AI draft is in the audit trail with the AI tool and model named. ## Where people decide - Change controls, CAPAs and controlled documents are approved in review and approval cycles with an author, at least one reviewer and at least one approver, never the author signing their own record, each signing with email and password. - A CAPA plan needs at least one action and a defined effectiveness method and criterion before approval; retraining alone blocks it. - Q Agent proposes actions and waits for a person's approval; nothing is ever signed or closed by an agent. ## Built to - 21 CFR Parts 210 and 211 - 21 CFR Part 820 (QMSR) - ICH Q9(R1) and ICH Q10 - 21 CFR Part 11 and EU GMP Annex 11 - EU GMP Chapters 1 and 4 - Draft EU GMP Annex 22 for the AI ## White paper One quality system, one AI layer The case for an Agentic QMS [Download the white paper](https://qlogix.ai/whitepapers#agentic-qms) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Agentic QMS ### What is an eQMS? An electronic quality management system (eQMS) is software that runs a manufacturer's quality processes, such as change control, CAPA and document control, as controlled electronic records with an audit trail and electronic signatures. The QLogix Agentic QMS is an eQMS for pharma, biotech, CDMOs, 503B facilities and medical device manufacturers, on the same records as the QLogix laboratory, maintenance and validation modules. ### What is an agentic QMS? An agentic QMS is a quality management system in which AI agents work on the quality records themselves: assessing a change, reviewing a CAPA plan or a document draft, or working a goal across the system. In the QLogix Agentic QMS every agent action waits for a person's approval, records the agent creates start in review, and nothing is signed or closed by an agent. ### How is AI controlled in a GxP quality system? In QLogix the AI only drafts, and each draft quotes the write-up or the attached documents it relies on. People decide every disposition and sign every approval with their email and password. Each AI draft, edit, decision and signature is an entry in the hash-chained audit trail with the AI tool and model named, so an inspector can see what the AI proposed and what a person decided. ### Does it support 21 CFR Part 11 and EU Annex 11? It is built to support both. The audit trail is append-only and chained with SHA-256, changes to existing values need a reason, and an electronic signature re-enters the email and password and records the meaning, the time, the record version and a hash of the content. Signing locks after five wrong passwords in 15 minutes, sessions end after 15 minutes idle, and each organization's records are kept separate by the database. ### Can we migrate from paper or another QMS? Controlled documents can be brought in: an administrator imports existing approved documents as their final approved PDFs, keeping the legacy document number, and signs the batch to attest that each file is the approved version. Each one becomes an effective controlled document, due for periodic review and revised like any other. The legacy import covers controlled documents; change controls and CAPAs are opened in QLogix. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Change control, CAPA and controlled documents are licensed together as the Agentic QMS. There is no card payment or self-serve plan. ## Reading - [AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writ…](https://qlogix.ai/blog/ai-document-management-validation-workflow) - [EU GMP Annex 22: what the draft AI annex asks of GMP manufacturersDraft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined i…](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft) - [FDA's seven-step credibility framework for AI, in plain termsFDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the dec…](https://qlogix.ai/blog/fda-ai-credibility-framework) ## Other modules - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) [All 22 product pages](https://qlogix.ai/products) --- # AI root cause investigation software for GMP | QLogix Source: https://qlogix.ai/products/root-cause-investigation # AI root cause investigation software that works with your existing QMS QLogix Root cause investigation is AI-enabled root cause analysis software for GMP manufacturers, sold on its own for organizations that already keep their deviations and CAPAs in another QMS. It takes a deviation, OOS or OOT result, complaint or audit finding from the first write-up to an approved, tested root cause and a corrective and preventive action plan. From the write-up and the attached SOPs, batch records and lab records, the AI drafts the problem statement with IS / IS NOT, the containment, 5 Whys chains, a fishbone and the CAPA plan, and every quote names its source. The investigator tests each fishbone cause as a hypothesis and dispositions it, and the conclusion is judged against eleven robustness checks before the record can be routed for signature. Each accepted action records the CAPA number it was raised under in your QMS, and the approved report goes back to the event there as a locked PDF. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the product does - Investigations of deviations (numbered in the DEV series), OOS or OOT results, complaints, audit or inspection findings and other events, opened from a write-up with the SOP, batch record, lab record, equipment or maintenance log, training record, specification or previous investigation attached. - Each attached document fingerprinted (SHA-256) and numbered D1, D2…; every quote in the analysis names the write-up or the document it comes from, and a withdrawn document stays on the record with the reason. - An AI-drafted root cause analysis in seven steps: the problem statement with IS / IS NOT for what, where, when and extent; containment with product and patient impact, other batches and the regulatory assessment; 5 Whys chains with each step marked Evidence or Assumption; a fishbone across six categories; the root cause; the conclusion; and the CAPA plan. - Every fishbone cause written as a testable hypothesis with a test method and what would support or refute it; the investigator records the result found, the evidence and the outcome, and the cause's disposition (root cause, contributing or ruled out) must agree with the outcome. - Statistical tests on the data behind a cause: Welch, paired or one-sample t, Mann-Whitney, ANOVA, Fisher's exact or chi-square, with the p-value, effect size and confidence interval recomputed on the server when the test is saved to the record. - A conclusion judged against eleven checks, among them every hypothesis tested, alternatives excluded, every IS and IS NOT explained, the cause verified and a system cause rather than human error, with a verdict of Robust, Adequate, Weak or Not established that sets how far the root cause confidence can go. - A CAPA plan of corrections, corrective actions and preventive actions with owners, timing and effectiveness checks, each accepted or rejected by the investigator, who can add their own. - Each accepted action records where it was raised in your own QMS (the system and the CAPA number); the reference is in the audit trail and on the report, and the panel Hand off to your QMS counts the actions raised. - A locked PDF report of the approved investigation, with the source and write-up, the analysis, the hypothesis tests, the conclusion and each action with its QMS reference, to attach to the event in your QMS. ## How it works beside your QMS Nothing in the investigation depends on the rest of QLogix. Open it with the event's number in your QMS (the deviation, complaint or OOS number) and the write-up, run the investigation with the AI and the checks, raise the accepted actions as CAPAs in your QMS and record their numbers, then route the investigation for review and approval and attach the approved report to the event. Your QMS keeps the deviation and the CAPAs; QLogix keeps a signed, audited investigation with every quote sourced. Licensed alone as the Root cause investigation suite (the investigations, document management for your SOPs and the regulatory library), or with CAPA management in the Agentic QMS, where the approved investigation opens a CAPA record prefilled with the root cause and the accepted actions. ## How it fits the platform Investigations open from wherever a problem is found. A complaint routed to investigation starts one with its text, summary and problem statement already written. An out-of-specification result in the LIMS opens an OOS investigation with the result, the determinations, the instruments, the standards and the analyst filled in, linked to the sample both ways, and the result is invalidated only on an assignable cause that investigation found. A CMMS work order that finds a failure opens a deviation investigation with the findings written, an out-of-tolerance calibration opens one with the impact assessment attached, and a deviation on an MES batch names the investigation that handles it. ## Where people decide The AI drafts the problem statement, the 5 Whys, the fishbone and the CAPA plan, reads each hypothesis result and can give a second opinion on the conclusion that changes nothing. The investigator tests each hypothesis, dispositions each cause with a reason, sets the root cause confidence, and accepts or rejects each action; reviewers and approvers sign the investigation with their password. The AI never dispositions a cause, accepts an action or signs. ## Where people decide - An investigation cannot be routed for review while a fishbone cause is open, a CAPA action is neither accepted nor rejected, or a disposition contradicts its hypothesis test. - Root cause confidence can be Confirmed only when the root cause's own test supports it. - Only an accepted action records where it was raised in your QMS; changing that reference needs a reason, and a signed investigation is changed only after it is reopened. ## Built to - 21 CFR 211.192 and 211.100(b) - FDA guidance on investigating OOS test results - 21 CFR Part 820 (QMSR) and ISO 13485:2016 clause 8.5 - EU GMP Chapter 1 and Chapter 8 - ICH Q9(R1) and ICH Q10 - 21 CFR Part 11 and EU GMP Annex 11 ## White paper From deviation to defensible root cause AI-enabled investigations for GxP manufacturing [Download the white paper](https://qlogix.ai/whitepapers#ai-enabled-investigations) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Root cause investigation ### What is root cause investigation software? Software that takes a quality event from the first write-up to a root cause that is shown, not assumed: a problem statement, the possible causes, the evidence and tests that rule each in or out, and the actions that remove the cause. In QLogix the AI drafts the analysis from the write-up and your documents, quoting its sources; the investigator tests and decides, and the investigation is signed. ### Can we use it with our existing QMS? Yes, it is sold on its own for that. Enter the event's number from your QMS when you open the investigation, record the CAPA number each accepted action was raised under in your QMS, and attach the approved report (a locked PDF) to the event there. QLogix does not need to hold your deviations or CAPAs. ### How does AI help with root cause analysis? The AI drafts the problem statement with IS / IS NOT, the 5 Whys chains and a fishbone from the write-up and the attached documents, quoting the source of each statement. Each fishbone cause comes as a hypothesis with a test plan; the investigator runs the tests and decides. The AI can review the conclusion as a second opinion, but it changes nothing and never dispositions a cause or signs. ### Does it handle OOS investigations? Yes. An investigation can be of an OOS or OOT result. With the QLogix LIMS, an out-of-specification result opens one with the result, determinations, instruments, standards and analyst filled in, and the result is invalidated only when that investigation finds an assignable cause. ### Does it support 21 CFR Part 11 electronic signatures? It is built for them. Investigations are routed for review and approval with electronic signatures, and each signature asks for the signer's email and password again. Every AI draft, edit with its reason, decision and signature is an entry in a hash-chained audit trail, and the approved record is locked. ### How is it licensed? Per user per month, alone as the Root cause investigation suite or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI in deviation investigations: faster, more consistent root cause analysisAI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable …](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When th…](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Other modules - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) [All 22 product pages](https://qlogix.ai/products) --- # CAPA software with effectiveness checks | QLogix Source: https://qlogix.ai/products/deviation-capa-management # CAPA software for GMP manufacturing: corrective and preventive actions closed on proven effectiveness QLogix CAPA management is CAPA software in the Agentic QMS that takes a corrective and preventive action plan from the problem that started it to the check that shows it worked. A CAPA opens from a root cause investigation with the problem statement, root cause, classification and accepted actions carried in, or from a complaint, an audit finding, a supplier issue, a signal, a change control, a batch record review, a management review or a risk assessment. The AI reviews the plan for adequacy, gaps and extent of condition; plan approval is refused while retraining is the only corrective or preventive action; each action is completed with evidence; and the CAPA closes only on a recorded effectiveness result. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the product does - CAPA records opened from an approved root cause investigation with the problem statement, root cause, classification, accepted actions and the effectiveness check carried in, or from a complaint, audit or inspection finding, supplier quality issue, signal or trend, change control, batch record review, management review or risk assessment. - An AI review of the plan for adequacy, gaps and extent of condition; plan approval refused while retraining is the only corrective or preventive action. - Actions with owners and due dates, completed with evidence, and an effectiveness check against the method and criterion defined in the plan. - Closure as Effective: close or Not effective: escalate, with the follow-up CAPA or investigation opened; the CAPA's stage (plan, plan approved, implementing, effectiveness check, closed) derived from what the record contains. ## How it fits the platform Root cause investigation is a separate product: the approved investigation opens the CAPA record or raises a change control, and an organization that keeps its CAPAs elsewhere can run the investigations alone. A change control whose effectiveness is not met cannot close until a CAPA is raised and linked. Backlog triage ranks the open deviations and CAPAs by risk, the work queue shows what each record is waiting for, and built-in alerts flag CAPA work past its target. ## Where people decide The AI reviews the plan and suggests an effectiveness method, criterion and timeframe. The CAPA owner records the evidence for each action and the effectiveness result, reviewers and approvers sign the plan and the closure, and the AI never closes a record or signs. ## Where people decide - A CAPA plan needs at least one action and a defined effectiveness method and criterion before approval; retraining alone blocks it. - Effective: close needs every action done and effectiveness met; Not effective: escalate needs effectiveness recorded as not met and leads to a follow-up CAPA or investigation. ## Built to - 21 CFR 211.192 and 211.100(b) - 21 CFR Part 820 (QMSR) and ISO 13485:2016 clause 8.5 - EU GMP Chapter 1 and Chapter 8 - ICH Q9(R1) and ICH Q10 CAPA system - 21 CFR Part 11 and EU GMP Annex 11 ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about CAPA management ### What is CAPA software? CAPA software records corrective and preventive actions from the problem that started them to the check that shows they worked. In QLogix a CAPA is opened from an approved investigation with the problem statement, root cause, classification and accepted actions carried in, reviewed by the AI for adequacy, approved, completed with evidence for each action and closed only on a recorded effectiveness result. Its stage is derived from what the record contains, not set by hand. ### How is CAPA effectiveness checked? The plan must define an effectiveness method and criterion before it can be approved, and the AI review suggests a method, criterion and timeframe. At closure, Effective: close needs every action done and effectiveness met, while Not effective: escalate needs effectiveness recorded as not met and leads to a follow-up CAPA or investigation. ### Is root cause investigation part of CAPA management? No, it is a separate product. The approved investigation opens a CAPA record prefilled with what it found, and an organization that keeps its CAPAs in another QMS can license Root cause investigation alone. ### Does it support 21 CFR Part 11 electronic signatures? It is built for them. CAPA plans and closures are routed for review and approval with electronic signatures, and each signature asks for the signer's email and password again. Every AI draft, edit with its reason, decision and signature is an entry in a hash-chained audit trail, and the approved record is locked. ### How is it licensed? As part of the Agentic QMS (with change control and document control), per user per month, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI-assisted CAPA: from root cause to a verified effectiveness checkA CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same caus…](https://qlogix.ai/blog/ai-capa-management-effectiveness) - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When th…](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Other modules - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) [All 22 product pages](https://qlogix.ai/products) --- # Complaint handling software for pharma and devices | QLogix Source: https://qlogix.ai/products/complaint-handling # Complaint handling software for pharma and medical devices: triage, adverse event check and investigation The QLogix complaint handling module is complaint handling software for pharmaceutical, biotech and medical device manufacturers that turns a complaint as received into a structured complaint record with quoted evidence, a route decided by a person and, where needed, a root cause investigation. The AI reads the complaint text and the attached email, letter or form and drafts the product, lot, event date, reporter, category, patient involvement and harm, each with the exact quote it rests on, and flags a possible adverse event for pharmacovigilance. The reviewer decides the route and confirms receipt to the reporter, and the preliminary evaluation drafts the problem statement, impact, hypotheses and next steps that start the investigation. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Intake from email, call center, web form, distributor, sales representative or other channels: the complaint text pasted as received with personal identifiers removed, the original attached as a PDF or Word document, or both, under a CMP number. - Supporting documents (correspondence with the reporter, a distributor's report, test data, photos of the defect) attached to the record, fingerprinted, audited and read by the AI with the complaint. - An AI-drafted complaint record: a summary, the product, lot or batch, event date, reporter, category, patient involved, harm reported and sample available, each with the exact quote it rests on, highlighted in the complaint text. - Eight complaint categories: product quality, container closure or packaging, labeling, lack of effect, patient reaction, device or delivery system, suspected counterfeit or tampering, and other. - An adverse event check: a possible adverse event is flagged with its rationale and a banner to route it to pharmacovigilance for assessment, with flags, missing information and suggested next steps listed for the reviewer. - The reviewer's route: route to pharmacovigilance, open an investigation, trend only, or not a complaint and close; the first decision needs no reason, a change does. - Confirmation of receipt emailed to the reporter from the record, prefilled with an AI acknowledgement that gives no medical advice and admits no fault, with the email logged on the record. - A preliminary investigation evaluation: the problem statement with IS / IS NOT, patient safety and product quality impact, other lots that could be affected, whether a regulatory or pharmacovigilance reporting assessment is needed, a suggested priority, hypotheses across the 6M categories and next steps with owner roles and timing. - The root cause investigation started from the complaint, with its text, summary, problem statement and the reviewer's note as the write-up, linked back to the complaint as its origin. ## How it fits the platform A complaint opened for investigation becomes an investigation in the Root cause investigation product, and the approved investigation opens the CAPA and, where needed, the change control. Connected records suggest other records that share the lot, product, equipment or cause wording, so a complaint on a lot already under deviation is seen next to it. Complaint rates can be trended in signal detection, and backlog triage ranks open complaints with the deviations and CAPAs by risk. Built-in alerts raise a possible adverse event without a decision as critical after 24 hours, and a complaint awaiting a decision after 5 days. Q Agent can log a complaint, open the investigation and draft the root cause analysis, each action waiting for a person's approval. ## Where people decide The AI drafts the complaint record, the adverse event flag, the acknowledgement and the preliminary evaluation; it never decides that an event is an adverse event, decides reportability or closes the record. The reviewer checks every field and edits with a reason, decides the route and the priority under the site's SOP, accepts or rejects each next step, and routes the complaint for review and approval with electronic signatures. ## Where people decide - The route (pharmacovigilance, investigation, trend only or close) is chosen by the reviewer; changing it needs a reason. - Adverse event determination and reportability belong to pharmacovigilance and regulatory affairs; the AI says only that an event is possible, and why. - The suggested priority is a suggestion; classification is the reviewer's decision under the site's SOP. ## Built to - 21 CFR 211.198 complaint files - 21 CFR Part 820 (QMSR) and ISO 13485:2016 clause 8.2.2 - EU GMP Chapter 8 complaints, quality defects and recalls - ICH Q10 pharmaceutical quality system - 21 CFR Part 11 and EU GMP Annex 11 ## White paper From complaint to closed CAPA, with nothing left behind An Agentic QMS for complaints, investigations and backlog [Download the white paper](https://qlogix.ai/whitepapers#complaints-investigations-backlog) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Complaint handling ### What is complaint handling software? Complaint handling software records each product complaint from receipt to its decision and, where needed, its investigation. In QLogix a complaint is an audited record with the complaint as received, the fields drafted by the AI with their quotes, the reviewer's route, the reply to the reporter and the linked investigation, signed in review and approval. ### How does the AI triage a complaint? It reads the complaint text and the attached documents and drafts the record fields, each with the exact quote it rests on, or "not stated" where nothing supports a value. It suggests a category, flags a possible adverse event with its rationale, and lists flags, missing information and next steps. Every field can be edited with a reason, and the earlier draft stays in the audit trail. ### Does it decide whether a complaint is reportable as an MDR or a field alert? No. Adverse event determination and reportability belong to the reviewer, pharmacovigilance and regulatory affairs. The AI flags a possible adverse event, and the preliminary evaluation says whether a regulatory or pharmacovigilance reporting assessment is needed and why, without stating that anything is or is not reportable. A possible adverse event without a decision raises a critical alert after 24 hours. ### Does it support 21 CFR 211.198 and 820.198 complaint files? It is designed to support them: each complaint is an audited record with the complaint as received, the product and lot, the nature of the complaint, the decision whether to investigate, the reply to the reporter and the linked investigation. For medical devices, the QMSR replaced 820.198 on 2 February 2026 with ISO 13485:2016 clause 8.2.2 and the complaint records of 21 CFR 820.35. ### How does a complaint become an investigation? When the reviewer decides Open an investigation, Start the root cause investigation opens one with the complaint text, summary, problem statement and the reviewer's note as its write-up, linked back to the complaint. The preliminary evaluation's hypotheses and accepted next steps give the investigator a starting point. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI-assisted CAPA: from root cause to a verified effectiveness checkA CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same caus…](https://qlogix.ai/blog/ai-capa-management-effectiveness) ## Other modules - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) [All 22 product pages](https://qlogix.ai/products) --- # Batch record review software: review by exception | QLogix Source: https://qlogix.ai/products/batch-record-review # Batch record review software: executed batch records reviewed by exception The QLogix batch record review module is batch record review software that reviews an executed batch record by exception and lists only the entries that do not meet their instruction, limit or good documentation practice. The AI checks every entry it can read in an electronic batch record export, a scanned paper batch record or a text extract against the master batch record and limits attached, recomputes calculations and reconciliations, and reports its coverage section by section, so the reviewer knows what was and was not checked. The reviewer dispositions every exception, links the deviations, adds anything the AI missed and signs a recommendation for release, hold, return or rejection; the batch disposition itself is decided and recorded in the site's QMS. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - A review per lot in the BRR series: the record type (electronic batch record export, scanned paper batch record, batch record extract or other), the lot or batch number, the product and the stage or line. - The executed batch record attached with the master batch record, specifications or limits, deviation or change records and equipment or environmental logs, or an extract of up to 24,000 characters pasted. - Each readable entry checked for values against limits; required entries, signatures, dates and second-person verifications; corrections initialled, dated and explained; calculations and reconciliations recomputed; sequence and holding times; component lots, equipment IDs and labels; and a deviation reference for every out-of-limit value or missed step. - Coverage reported section by section as reviewed, partly reviewed or not reviewed, with what could not be read; illegible entries are reported, never guessed, and a missing limit is listed rather than invented. - Only the exceptions listed (E1, E2…), each with the step, what was recorded against what was required, the exact quote and its location, a category, a severity of Critical, Major or Minor with the rationale and a suggested action; entries that passed are counted. - Exception categories: out of limit, missing entry, missing signature or verification, late or unexplained correction, calculation error, sequence or timing, equipment, material or label, documentation practice and other. - Each exception dispositioned with a reason: link deviation with its reference, accept with justification, return for correction or not an exception; exceptions the reviewer finds are added to the same list. - A recommendation of release, hold pending deviation closure, return for correction or rejection, routed for review and approval with electronic signatures. ## How it fits the platform An exception linked to a deviation carries the deviation's reference, and the deviation is investigated in the Root cause investigation product, where it can lead to a CAPA that names batch record review as its origin. Connected records link the review to the deviations, investigations and CAPAs, and suggest other records that share the lot or the equipment. For batches executed in the QLogix MES, each step is executed and verified with signatures as it happens, entries outside their limits are flagged on the batch, and QA reviews and releases the batch there. Built-in alerts flag batch review exceptions left open after 3 days and a batch on hold or returned after 7 days, and the work queue shows each review's stage and what it is waiting for. ## Where people decide The AI checks the entries and lists the exceptions with their quotes and a suggested action; it never dispositions an exception, records the recommendation or releases a batch. The reviewer dispositions every exception with a reason, adds what the AI missed and records the recommendation; the reviewers and approvers sign it. The batch disposition is decided and recorded in the site's QMS by the people authorized to release. ## Where people decide - Every exception is dispositioned by the reviewer with a reason; nothing is deleted, and an exception the AI missed is added by hand. - A review cannot be routed until the record has been reviewed, every exception is dispositioned, every linked deviation has its reference and the recommendation is recorded. - The recommendation is signed here; the batch disposition is decided and recorded in the site's QMS. ## Built to - 21 CFR 211.186 and 211.188 master and batch production records - 21 CFR 211.192 production record review - EU GMP Chapter 4 documentation - 21 CFR Part 11 and EU GMP Annex 11 - ALCOA+ data integrity principles ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Batch record review ### What is review by exception for batch records? Review by exception means the reviewer reads what does not meet its requirement instead of every entry that does. The AI checks each readable entry against its instruction, limit or good documentation practice and lists only the exceptions, with the quote and its location. Entries that passed are counted, and coverage is reported section by section so the reviewer sees what was not checked. ### What does the AI check in an executed batch record? Values against their limits, missing entries, signatures and second-person verifications, corrections without a date, reason or initials, calculations and reconciliations such as yield and labels, sequence and holding times, the identity of component lots, equipment and labels, and a deviation reference for every out-of-limit value or missed step. It works only from the documents provided: when a limit is not available, it says so and lists the missing limit for the reviewer. ### Can it review scanned paper batch records? Yes. A scanned paper batch record, an electronic batch record export or a text extract can be reviewed. Pages or entries that cannot be read are reported as partly reviewed or not reviewed with a note; the AI does not guess at illegible entries. ### Does it release the batch? No. The reviewer records a signed recommendation: release, hold pending deviation closure, return for correction or rejection. Batch disposition is decided and recorded in the site's QMS, and the AI never releases a batch. ### Is the review signed under 21 CFR Part 11? The recommendation is routed for review and approval with electronic signatures that ask for the signer's email and password again. Every AI review, disposition, added exception and signature is an entry in a hash-chained audit trail, and the approved record is locked. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing mode…](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) ## Other modules - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) [All 22 product pages](https://qlogix.ai/products) --- # GMP change management software with AI assessment | QLogix Source: https://qlogix.ai/products/change-control # Change management software for GMP: documents, equipment and facilities under change control QLogix change management is GMP change control software that takes a proposed change to a controlled document, a piece of equipment, a facility area or a utility from request to closure on audited, signed records. The change lists what it touches, each item with its current and proposed state. The AI assesses the change as soon as it is proposed: its classification, its impact in nine areas from product quality to data integrity, the ICH Q9 risk, the GMP requirements each item must meet before it is used (document revision, approval and training; requalification, calibration and maintenance; cleanroom, HVAC and utility qualification) with their regulatory basis, the documents and training affected, the tasks and approvers, the effectiveness check and whether regulatory screening or validation is needed, each point quoting its source. People plan the tasks, own each function's part and sign it off, approve the change to implement, meet every requirement with evidence, complete the tasks and close the change once the effectiveness check is recorded. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - A change proposed with its type (manufacturing process, equipment or facility, computerized system, component or supplier, specification or method, document or procedure, packaging or labelling, organization or outsourcing, other), urgency (Routine, Expedited or Emergency, implemented before approval), the products, sites or lines affected, the description and the reason. - The scope of the change, item by item: each controlled document (by document number), piece of equipment (by asset tag) and facility area or utility (by room or system) the change touches, with its current state, its proposed state and the QLogix record that carries it. The requester lists what they know; the AI adds what the description and documents name; a document or equipment change names its scope before it goes for approval. - Supporting documents attached and read by the AI: the change request, the current procedure, drawings or URS, validation documents, supplier documentation, or an existing impact or risk assessment. - A change impact assessment by the AI: Major, Minor or Like-for-like with the rationale and quoted source, and the impact on product quality, patient safety, validation and qualification, regulatory filings, documents and procedures, training, supplier and materials, data integrity and computerized systems, and environment, health and safety. - The ICH Q9 risk with severity, probability, detectability and overall level, and whether regulatory screening and validation are needed; the documents and training affected, suggested tasks with timing, approver roles, an effectiveness check and questions for the requester. - The GMP requirements per scope item, each with its regulatory basis cited. Documents: the new version reviewed, approved and effective under document control, training recorded before the effective date, the superseded version withdrawn, forms and referencing documents updated (EU GMP Chapter 4, 21 CFR 211.100 and 211.188). Equipment: the impact on the qualified state and the requalification scope, calibration, the preventive maintenance plan and asset record, cleaning validation, drawings, logbooks and operating procedures (EU GMP Chapter 3 and Annex 15, 21 CFR 211.63 to 211.68). Facilities and utilities: cleanroom classification and requalification, HVAC balancing and pressure cascade, environmental monitoring, utility qualification, as-built drawings, cleaning and disinfection, flows and the contamination control strategy (EU GMP Chapter 3 and Annex 1, 21 CFR 211.42 to 211.48). Each requirement is met with evidence or recorded as not applicable with a reason; closure waits for every one. - Assess again with a reason: the new assessment replaces the open suggestions and keeps the edits, decisions and completed items. - Up to 40 tasks by type and timing (before approval, before implementation, at implementation, within 30 or 90 days after), each done with evidence of completion or marked not required with a reason. - Tasks grouped by function (QA, QC, Engineering and the others), each function with a named owner who is emailed the tasks and signs that the functional assessment is complete. - Approval to implement as a signed milestone, then an effectiveness check recorded as met, not met or not required; a change not met cannot close until a CAPA is raised and linked. - Closure as the final signature, which locks the record, once every task is done, every GMP requirement is met or not applicable and the effectiveness result is recorded; a change can instead be recorded as Rejected or Withdrawn with a reason. ## How it fits the platform A change control can be raised from an investigation or a CAPA and stays linked to it, so the action and its origin read as one file. When the assessment says a regulatory filing may be needed, a regulatory impact screening in Regulatory submissions assesses the registered details affected and suggests a reporting route per market, from a Prior Approval Supplement, CBE-30, CBE-0 or Annual Report in the United States to Type IA, IB or II variations in the EU and UK; regulatory affairs confirms or changes each route. An alert is raised when a change screened as likely to need prior approval has no RA decision after seven days. The procedure the change requires is revised in document control, a change to a qualified asset links to the validation project that requalifies it, and a CMMS modification or decommissioning runs under the change control. The implementation can be planned as a project in Project management, Backlog triage ranks open change controls by risk, and the Agentic QMS overview lists emergency and expedited changes still open. ## Where people decide The AI assesses the change and suggests the tasks, approvers and effectiveness check; the author edits the plan, the function owners complete their tasks and sign off, the approvers sign the approval to implement, regulatory affairs decides each reporting route, and the closure is signed when the effectiveness result is on record. The AI's classification and risk level are drafts until a person accepts or changes them with a reason. ## Where people decide - Approval to implement needs an assessment and at least one task; the final approver's signature records the decision Approved, and the author cannot approve their own change. - Every function with tasks has a named owner who signs that the functional assessment is complete before closure. - Closure is blocked while any task or GMP requirement is open, while the effectiveness result is missing, or while it is not met and no CAPA is linked. ## Built to - ICH Q9(R1) quality risk management - ICH Q10 change management system - 21 CFR 211.42 to 211.68, 211.100 and 21 CFR 314.70 - 21 CFR Part 820 (QMSR) - EU GMP Chapters 1, 3 and 4, Annex 1 and Annex 15 - 21 CFR Part 11 and EU GMP Annex 11 ## White paper From change control to filed dossier AI-enabled regulatory submissions on an Agentic QMS [Download the white paper](https://qlogix.ai/whitepapers#regulatory-submissions) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Change management ### What is change control software? Change control software records a proposed change to a GMP process, system, material or document, assesses its impact and risk, and controls its approval, implementation and closure with an audit trail and electronic signatures. QLogix change control does this on the same records as CAPA, document control and the other QLogix modules, with an AI that drafts the impact assessment. ### How does the AI assess the impact of a change? It reads the change description and the attached documents and classifies the change as Major, Minor or Like-for-like, rates the impact in nine areas as none identified, possible or likely, and scores the ICH Q9 risk by severity, probability and detectability. Each point quotes its source, and the author accepts it or edits it with a reason. ### How is the regulatory impact of a change assessed? When the assessment says a filing may be needed, a regulatory impact screening is opened in Regulatory submissions and linked to the change. It lists the registered details affected and suggests a reporting route for each market where the product is registered; regulatory affairs confirms the route, chooses a different category or records no filing. The filing itself is made in your regulatory system. ### How are equipment and facility changes handled under GMP? The change lists each piece of equipment, by asset tag, and each facility area or utility, by room or system, with its current and proposed state. The AI sets out the GMP requirements each must meet before use: for equipment, the impact on the qualified state and the requalification scope, calibration of the instruments involved, the preventive maintenance plan and asset record, cleaning validation, drawings, logbooks and operating procedures; for a facility or utility, cleanroom classification and requalification, HVAC balancing and the pressure cascade, environmental monitoring, utility qualification and monitoring, as-built drawings, cleaning and disinfection, flows and the contamination control strategy. Each requirement cites its basis in EU GMP and 21 CFR 211, is met with evidence or recorded as not applicable with a reason, and the change cannot close while one is open. A change to a qualified asset links to the validation project that requalifies it and to the CMMS asset. ### How are document changes handled? The change names each controlled document by its number with the revision proposed. The GMP requirements the AI sets out follow document control: the new version reviewed and approved, an effective date, training of the affected roles recorded before that date, the superseded version withdrawn from points of use, and the forms and referencing documents updated. The revision itself is made in document control, where the new version carries the change control's reference, and the change closes once the version is effective and the requirement is met with that evidence. ### How are emergency changes handled? The change is recorded with the urgency Emergency, implemented before approval, and still goes through its assessment, a decision in review and approval, and closure. The Agentic QMS overview lists emergency and expedited changes that are still open. ### What happens when the effectiveness check is not met? The change cannot close. A CAPA is raised and linked to the change control, and closure is possible once it is, so the follow-up is on record with the change that needed it. ### How is it licensed? Change control is part of the Agentic QMS, with CAPA and controlled documents, licensed per user per month on its own or with other modules and ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI-assisted CAPA: from root cause to a verified effectiveness checkA CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same caus…](https://qlogix.ai/blog/ai-capa-management-effectiveness) - [AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writ…](https://qlogix.ai/blog/ai-document-management-validation-workflow) ## Other modules - [Document controlDocument control software for GMP: AI review of each SOP draft, signed approval on an effective date, read-and-understood training signatures and periodic review.](https://qlogix.ai/products/document-control) - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) [All 22 product pages](https://qlogix.ai/products) --- # Document control software for GMP SOPs and training | QLogix Source: https://qlogix.ai/products/document-control # Document control software for GMP: SOPs, controlled documents and training QLogix document control is GMP document management software for controlled documents: SOPs, work instructions, policies, forms, specifications, protocols, reports and master batch records, each with its versions, owner and review period. The AI reviews each draft version for structure, clarity, completeness, consistency, compliance and data integrity, and says whether training is required and for which roles. The author dispositions every finding, approvers sign, the version becomes effective on its date, and each trained person signs that they have read and understood it. Periodic review falls due from the effective date, and every exported copy is stamped with its state. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Controlled documents numbered from the series of their type (SOP, WI, POL, FRM, SPEC, PRT, RPT, MBR or DOC), with the title, department, document owner, what it supersedes, purpose, keywords and a periodic review every 12, 24 or 36 months. - Versions, each a fingerprinted file: a revision is added as the next version, starts as a draft and goes through review and approval again, and the previous effective version becomes superseded. - An AI review of each draft: the standard structure (purpose, scope, responsibilities, definitions, procedure, records and forms, references, revision history) and findings by severity (Major, Minor, Observation) and category, each with its location and a suggested fix. - The training impact from the same review: whether people must be trained before the version becomes effective and the roles to train, with the regulations the document implements and questions for the approvers. - Findings dispositioned with a reason as resolved in the draft, not applicable or open; every Major finding must be dispositioned before approval. - Approval in review and approval with electronic signatures, set to make the version effective on its effective date or to make the document obsolete. - Read-and-understood training: each trained person signs "I have read and understood version X" with email and password, once per person and version. - Periodic review due from the effective date, shown in the work queue and the Agentic QMS overview; an obsoleted document is withdrawn from use and no review falls due. - Stamped copies: CONTROLLED COPY - APPROVED, IN REVIEW - NOT FOR USE, DRAFT - NOT FOR USE or SUPERSEDED - FOR REFERENCE ONLY, exported as locked PDFs that cannot be edited. ## How it fits the platform Document management puts every document the quality system produces in one searchable index: controlled documents and their versions, validation protocols and reports, risk assessments, change controls, CAPAs, investigations, QC specifications and certificates of analysis, with every attached file and its fingerprint. The hub stores nothing of its own, so what it finds is the record itself, with its signatures and audit trail. Existing approved documents come in through the legacy import as their final approved PDFs, keeping their document numbers. A change control names the documents and training it affects, and the revision is made here. The electronic lab notebook lets an analyst follow only the effective version of a method, and the controlled document lists the runs that followed it. When an asset is decommissioned in the CMMS, the close-out lists the controlled documents to obsolete. ## Where people decide The AI reviews the draft and suggests wording; it never states that a document is approved or compliant. The author decides what to change and dispositions each finding with a reason, the approvers sign the version effective or obsolete, and each trained person signs their own read-and-understood attestation. ## Where people decide - A version becomes effective only through approval in review and approval, on the effective date set before approval. - Every Major finding from the AI review is dispositioned with a reason before the version can be approved, and the roles to train are named when training is required. - Read and understood is an electronic signature, with email and password, once per person and version. ## Built to - 21 CFR 211.100 written procedures - 21 CFR Part 820 (QMSR) and ISO 13485 clause 4.2.4 - EU GMP Chapter 4 Documentation - 21 CFR 211.25 and EU GMP Chapter 2 training - 21 CFR Part 11 and EU GMP Annex 11 - MHRA GxP data integrity guidance ## White paper One quality system, one AI layer The case for an Agentic QMS [Download the white paper](https://qlogix.ai/whitepapers#agentic-qms) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Document control ### What is document control software? Document control software manages controlled documents such as SOPs, work instructions and specifications through drafting, review, approval, distribution, training, periodic review and obsolescence, with an audit trail and electronic signatures. QLogix document control does this on the same audited records as change control, CAPA and the other QLogix modules, with an AI review of each draft. ### How does the AI review an SOP? It checks the draft against the standard structure, reads the procedure as the person who must follow it, and compares it with the referenced or linked documents provided. Findings are listed by severity and category with their location and a suggested fix, together with the training impact. The author dispositions each finding with a reason; the AI does not approve anything. ### How is read-and-understood training recorded? After the version is approved, each trained person signs "I have read and understood version X" with their sign-in email and password. It is an electronic signature recorded once per person and version, with the time, in the audit trail. ### How does periodic review work? Each document has a review period of 12, 24 or 36 months, and the next review falls due that long after the effective date. Documents due for periodic review appear in the work queue and on the Agentic QMS overview; an obsoleted document has no review due. ### Can we import our existing SOPs? Yes. An administrator imports existing approved documents as their final approved PDFs, giving for each the legacy document number, title, type, version, effective date and who approved it and when, and signs the batch. A file that is not a PDF, or a document number already in use, is refused, and every import is audited with the legacy approval details. ### How is it licensed? Controlled documents are part of the Agentic QMS, with change control and CAPA; Document management, the index of every QMS document, is licensed as its own module. Both are per user per month, on their own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writ…](https://qlogix.ai/blog/ai-document-management-validation-workflow) ## Other modules - [Risk assessmentQuality risk management software for FMEA under ICH Q9(R1): design, process, equipment and CSV templates, RPN thresholds, mitigation tracked to residual risk.](https://qlogix.ai/products/quality-risk-management-fmea) - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) [All 22 product pages](https://qlogix.ai/products) --- # Quality risk management software and FMEA for pharma | QLogix Source: https://qlogix.ai/products/quality-risk-management-fmea # Quality risk management software for FMEA under ICH Q9(R1) The QLogix risk assessment module is quality risk management software that runs each failure mode and effects analysis (FMEA) as one audited, signed record, from the first draft to the acceptance of the residual risk. An assessment starts from a design, process, equipment or CSV FMEA template; the AI drafts the register from the description of the subject, the attached documents and the linked records, and the team scores severity, occurrence and detection. RPN thresholds from the organization's risk management procedure class every risk Critical, High, Medium or Low and fix its follow-up. Critical and High risks carry mitigation actions with owners, due dates and expected scores, are reassessed from evidence once implemented, and the residual risk is accepted by QA with a justification or sent back for further action. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Four templates: design FMEA (dFMEA) for a product, formulation, container closure or device design; process FMEA (pFMEA) for a manufacturing, packaging or laboratory process; equipment FMEA for equipment, a facility or a utility; and CSV FMEA, the GAMP 5 functional risk assessment of a computerized system. - Guidance in each template for the item, failure mode, effect, cause and controls, the sources the AI reads (batch records and deviations, maintenance and calibration records, the URS and supplier assessment) and a worked example on the form. - An AI-drafted register: the summary, method, assumptions and questions for the team, and 8 to 150 failure modes scored 1 to 10 for severity, occurrence and detection, with a mitigation action, owner role, due date and expected scores for every Critical and High risk. - The risk priority number (RPN = S × O × D) classed against thresholds set from your procedure; by default Critical at an RPN of 200 or a severity of 9 and above, High at 100, Medium at 50. Changing a threshold reclasses the whole register and needs a reason. - Rescoring, additions and removals each with a reason in the audit trail; a register view ranked by RPN with the Pareto, the severity by occurrence matrix and the rating scales. - Mitigation tracking: actions assigned to named people, who are emailed, with statuses from planned through implemented to verified effective, and evidence files (change control closure, test report, training record) fingerprinted and linked to the row. - Post-mitigation reassessment of severity, occurrence and detection from the evidence, with the residual RPN and class; a residual at or above the acceptable RPN needs further action or QA's acceptance with a justification. - The risk assessment report as a locked PDF with the register, follow-up, reassessment, signatures and audit trail; every export is logged. ## How it fits the platform The AI drafts from what is attached or linked to the assessment: batch records, deviations and specifications for a process FMEA, maintenance and calibration records for an equipment FMEA, the URS and supplier assessment of a computer system validation for a CSV FMEA. A mitigation that changes a process, a procedure or a piece of equipment is carried by a change control or CAPA linked to the assessment, and the reassessment rests on its evidence. Q Intelligence can also build a draft FMEA or PFMEA in the conversation from deviations, investigations and SOPs, ranked by RPN, for the team to review. The formal, controlled assessment is a risk assessment record, listed in document management's index with its signatures and audit trail. ## Where people decide The AI drafts the register and proposes scores and actions; the cross-functional team rescores with reasons, owners implement the actions and attach the evidence, and the assessment is approved with electronic signatures. QA accepts any residual risk that stays at or above the acceptable RPN, with a justification, and closure signs and locks the record. The AI never accepts a risk. ## Where people decide - Assessment approval needs every Critical and High risk to have an action, an owner, a due date and expected scores, with a justification where the expected residual is not acceptable. - A residual risk at or above the acceptable RPN cannot close without QA's acceptance and justification, or further action and a new reassessment. - Closure needs every planned action done with evidence or cancelled and Critical risks verified effective; approval signs and locks the record. ## Built to - ICH Q9(R1) quality risk management - GAMP 5 (Second Edition) Appendix M3 functional risk assessment - EU GMP Chapters 3 and 5 and Annex 15 - ISO 14971 for medical device design - AIAG and VDA FMEA Handbook - 21 CFR Part 11 and EU GMP Annex 11 ## White paper Risk assessments that stay true to the data Q Intelligence and the AI-enabled FMEA [Download the white paper](https://qlogix.ai/whitepapers#quality-risk-management-fmea) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Risk assessment ### What is FMEA in pharma? Failure mode and effects analysis (FMEA) lists how a process, product, piece of equipment or computerized system can fail, what the effect would be and why it would happen, then scores severity, occurrence and detection to rank the risks. ICH Q9(R1) names it among the tools of quality risk management. The ranking decides where controls and mitigation go first. ### What is a risk priority number? The product of the severity, occurrence and detection scores, each from 1 to 10. QLogix classes each risk against thresholds set from your risk management procedure. Because controls do not change severity, a failure mode at the critical severity keeps its class after mitigation and stays visible for review. ### What is the difference between a pFMEA and a dFMEA? A design FMEA examines a product, formulation, container closure or device design, function by function against its design requirements. A process FMEA examines the steps of a manufacturing, packaging or laboratory process against their parameters and limits, with causes grouped in the 6 Ms. QLogix also has equipment and CSV FMEA templates. ### How are mitigation actions tracked? Each action has an owner, a due date and the expected scores, and moves from planned to implemented, with evidence, and verified effective. Once it is implemented the row is reassessed from the evidence, and the residual RPN shows whether the risk is mitigated or needs QA's acceptance. ### Can the AI accept a risk? No. The AI drafts the register and proposes scores and actions; people rescore with reasons, implement and evidence the actions, and QA accepts any residual risk that is not acceptable, with a justification. Approval and closure are signed electronically. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Computer software assurance: what FDA's final guidance changes for validationFDA's final guidance on computer software assurance replaces test-everything validation with effort matched to risk. Features that could compromise safety get scripted testing; the…](https://qlogix.ai/blog/computer-software-assurance-final-guidance) ## Other modules - [Regulatory submissionsRegulatory submissions software: post-approval changes screened per market, NDA and BLA sections drafted on an eCTD template, every claim quoted from its source.](https://qlogix.ai/products/regulatory-submissions) - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) [All 22 product pages](https://qlogix.ai/products) --- # Regulatory submissions software: changes, NDA and BLA | QLogix Source: https://qlogix.ai/products/regulatory-submissions # Regulatory submissions software for post-approval changes and NDA and BLA dossiers The QLogix regulatory submissions module is regulatory submissions software that documents regulatory affairs' assessment of a change and drafts submission text in which every claim quotes its source. For a post-approval change, the AI screens the change control for regulatory impact: the registered details affected, the supporting data, and a suggested reporting route in each market where the product is registered, which regulatory affairs confirms or changes with a reason. For a new NDA or BLA, the AI checks the documents against an eCTD template, Modules 1 to 5, section by section, and drafts the sections regulatory affairs asks for. A person verifies or rejects each claim before a section can be approved, and the filing itself is made in the organization's regulatory system. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Change screening by type of change (manufacturing process or equipment, component, material or supplier, container closure, specification or analytical method, site, labelling, shelf life or storage) for the markets where the product is registered: United States, European Union, United Kingdom, Canada, Japan and others. - A regulatory impact assessment: what changes, the registered details affected by CTD section with the quote, a supporting-data checklist (provided, incomplete, missing), questions for regulatory affairs, and a classification: prior approval likely, notification likely sufficient, likely no filing, or cannot determine. - A suggested reporting route per market with the condition that puts the change there: Prior Approval Supplement, CBE-30, CBE-0 or Annual Report in the United States; Type IA, IAIN, IB or II variation in the EU and UK; Level I to IV in Canada; partial change approval or minor change notification in Japan. - Regulatory affairs' decision for each market, with a reason: suggested route confirmed, a different category, or no filing in this market. - Submission sections drafted by the AI (description of the change, justification, impact on quality, safety and efficacy, summary of supporting data, comparability, cover letter), each fact marked [C1], [C2] with the exact quote from its source and unsupported points listed as not sourced. - NDA 505(b)(1), NDA 505(b)(2) and BLA 351(a) applications on an eCTD template, Modules 1 to 5 with the US regional Module 1, each section with its purpose, contents, analytical expectations, regulatory basis and the deficiencies that most often lead to a refuse-to-file or an information request. - A readiness check of the dossier documents against every template section (ready to draft, partial, missing, not applicable) with the filing risks, and section drafting guided by the template and the regulatory library passages it names. - Claim-by-claim verification: a section is approved only when every claim is verified, approved sections cannot be edited, and the decision is recorded and signed in review and approval. ## How it fits the platform A change control's impact assessment says whether regulatory screening is needed, and a change that may need a filing goes to a regulatory impact screening, with the change control record, specifications, qualification and validation reports and stability data attached as its sources. The CMC facts a section states come from documents named in its claims, so the reviewer checks each one against its quote. Open submission records appear in the work queue with the rest of the quality system, and a prior-approval change without a regulatory affairs decision raises an alert after seven days. ## Where people decide The AI screens the change, suggests a category with its confidence and drafts the text; regulatory affairs decides each route, verifies or rejects each claim, approves each section and records the decision, which is signed in review and approval. The AI does not state that a filing is or is not required. The filing is made in the organization's regulatory system; the QLogix record documents the assessment, the sources and the decision. ## Where people decide - The reporting route in each market is decided by regulatory affairs with a reason; the AI's category is a suggestion. - A section is approved only when every claim in it is verified; an approved section cannot be edited. - Proceeding to submission needs at least one approved section, and the decision is signed in review and approval. ## Built to - 21 CFR 314.70 and 601.12 post-approval changes - 21 CFR 314.50 and 601.2 applications - ICH M4 Common Technical Document - Regulation (EC) No 1234/2008 on variations - Health Canada Post-Notice of Compliance changes guidance - 21 CFR Part 11 and EU GMP Annex 11 ## White paper From change control to filed dossier AI-enabled regulatory submissions on an Agentic QMS [Download the white paper](https://qlogix.ai/whitepapers#regulatory-submissions) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Regulatory submissions ### What is a regulatory impact assessment? It decides whether a change to an approved product touches registered details in the dossier and, if it does, which reporting category applies in each market. In QLogix the AI lists the registered details affected with quotes, the supporting data provided or missing and a suggested route per market. Regulatory affairs confirms or changes each route with a reason. ### Which post-approval reporting categories does it consider? Prior Approval Supplement, CBE-30, CBE-0 and Annual Report in the United States; Type IA, IAIN, IB and II variations in the EU and UK; Level I to IV changes in Canada; and partial change approval or minor change notification in Japan. Each suggestion names the condition that puts the change in its category. ### Does it draft NDA and BLA dossiers? It supports NDA 505(b)(1), NDA 505(b)(2) and BLA 351(a) applications on an eCTD template covering Modules 1 to 5 with the US regional Module 1. The AI checks the documents against every template section, then drafts the sections regulatory affairs chooses, each claim quoting its source. Forms, datasets and study reports are compiled, not drafted. ### Does QLogix file the submission or publish the eCTD? No. The filing is made in your regulatory system. The QLogix record documents the impact assessment, the sources, the verified sections and regulatory affairs' signed decision. ### How is the AI kept from inventing data in a submission? Every statement of fact carries a claim marker with the exact quote from the change description or a named document, and points the sources do not support are listed as not sourced instead of written. A person verifies or rejects each claim, and a section cannot be approved until every claim is verified. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Other modules - [LIMSLIMS for pharma QC laboratories: approved specifications, results judged by the system, OOS under investigation, calibration lockouts and a signed CoA.](https://qlogix.ai/products/lims) - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) [All 22 product pages](https://qlogix.ai/products) --- # LIMS for pharma QC labs with OOS control and CoA | QLogix Source: https://qlogix.ai/products/lims # LIMS for pharma QC laboratories: specifications, results, OOS and certificates of analysis The QLogix LIMS is a laboratory information management system that records every in-process, release, stability and raw material result against an approved specification, on instruments in calibration and reference standards in date. The system calculates each result from the determinations as read and judges it against the limits when it is saved, so an out-of-specification or out-of-trend result is identified at entry, not at review. An out-of-specification result stands until an investigation linked to the sample finds an assignable cause; a passing repeat never replaces it. A second person who recorded no result reviews the sample, QA approves it, and the certificate of analysis is produced from the approved record with the signatures. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Specifications per material and testing stage (in-process control, release, stability, raw material acceptance, investigation or retest), with each test's method, unit, reported decimals, determinations, limits, alert limits, maximum %RSD and limit of quantitation; tests drafted by the AI from the registered specification, monograph or method are marked as such and checked one by one. - Specification approval with electronic signatures and revision as a new version: the current version stays in force until the next is approved, and samples keep the version they were registered under. - Samples registered against the specification in force, with lot, sampling point, quantity, storage, due date and priority; stability samples carry the study, condition and time point; results are recorded only on a sample whose receipt is on the chain of custody. - Chain of custody from receipt: the receiver checks identity, containers, seal, temperature and quantity and signs electronically with the date and time; every later handover, to an analyst or a contract laboratory, is signed too. - Sample inventory: each container with exactly where it is stored (building, room, storage unit, shelf, box, position) and its balance, computed from every use an analyst records in mL or µL (or g, mg, units) against the test, every move, aliquot, signed loss and signed disposal; results cannot go for review until every analysis is accounted for. - Analyst certification on test methods and analyzers, with certification and recertification dates and signed certify, recertify, suspend and withdraw steps: results are entered and reviewed only by people certified on the method and the analyzer, and recertifications coming due raise alerts. - Barcodes: sample, container, standard and instrument labels with 1D (Code 128, GS1-128, Code 39) and 2D (Data Matrix, QR, PDF417) symbols, printed from the browser or as ZPL for Zebra printers, and read with a handheld scanner, the camera or a photo. - Results entered as read under the analyst's own account with the system time; the analysis time checked, a late entry justified, and raw data files (chromatograms, printouts) fingerprinted with SHA-256 and linked to each result. - Calculation in exact decimal arithmetic with rounding half up to the reported decimals (USP General Notices 7.20) before the comparison with the limits; %RSD, individual determinations and values below the limit of quantitation handled by rule; out-of-trend results need a written evaluation. - Instrument control: calibration and qualification on each instrument, a result refused on an instrument out of service or out of calibration on the day of analysis, and a failed calibration that lists every sample tested on it since its last passing calibration. - Reference standards and reagents in quarantine until their certificate is checked and they are released; a result is refused on a standard that was not released or was past its expiry or retest date. - OOS management: an out-of-specification investigation opened from the sample with the result, determinations, instruments, standards and analyst filled in; invalidation only by that investigation's assignable cause; repeats with a reason; corrections only by the analyst who recorded the result, never into compliance. - Laboratory indicators, result trends as control charts with Nelson run rules and capability, CSV downloads and a results API; controlled interfaces for chromatography data systems, instrument software and the ERP, each validated before it is made active. ## How it fits the platform Investigations are built in: an OOS investigation opens from the sample, and a failed calibration opens a deviation investigation pre-filled with the instrument and every sample to assess. Results entered in the electronic lab notebook are posted to the LIMS sample with the run's reference, so the reviewer opens the work behind each result. A LIMS instrument can be linked to its calibration record and its maintenance asset in the CMMS, and a result is then accepted on it only while those records also say it is fit for use. The MES registers in-process and release samples of a lot in the LIMS against the specification in force, and the batch is not dispositioned until release testing is approved. An instrument data system sends results with their raw data and reads the worklist; an ERP requests samples for a lot and reads the dispositions. A result received through an interface is checked exactly like a typed one and shows where it came from. ## Where people decide The AI drafts the tests from the specification document; the analyst checks them and the specification is approved with signatures. The analyst records what was read; the system calculates, judges and refuses what does not comply. A reviewer who recorded no result on the sample checks it against the raw data, and QA approves it with an electronic signature, at which point the system sets the conclusion and the certificate of analysis can be issued. ## Where people decide - Whether a numeric result conforms is decided by the server when it is saved, from the determinations and the limits; nobody types a pass. - An out-of-specification result is invalidated only by an assignable cause found in a linked investigation, and stays on the record struck through with the reason. - Nobody who recorded a result on a sample can review or approve it; QA approval sets the conclusion and releases the certificate of analysis. ## Built to - 21 CFR 211.160, 211.165 and 211.194 laboratory controls and records - 21 CFR 211.192 and FDA guidance on investigating OOS results - EU GMP Chapter 6 Quality Control - USP General Notices 7.20 rounding - 21 CFR Part 11 and EU GMP Annex 11 - ALCOA+ data integrity principles ## White paper Lab and plant, connected LIMS, calibration, logbooks and CMMS on one AI-enabled infrastructure [Download the white paper](https://qlogix.ai/whitepapers#connected-lab-and-plant) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about LIMS ### What is a LIMS for pharma? A LIMS (laboratory information management system) for pharma holds the QC laboratory's specifications, samples, results and approvals as GMP records. The QLogix LIMS records each result against an approved specification, checks the instruments and reference standards used at the time of analysis, judges the result, and routes the sample for second-person review and QA approval with electronic signatures. ### How does the LIMS handle out-of-specification (OOS) results? The result is judged out of specification when it is saved and shown in red on the sample. The OOS investigation opens from the sample with the result, determinations, instruments, standards and analyst filled in, linked both ways. Only an assignable cause found by that investigation invalidates the result, which stays on the record struck through, and a sample with an OOS result standing cannot be cancelled. ### Can the LIMS issue a certificate of analysis? Yes, once QA has approved the sample. The certificate of analysis is produced from the approved record with the signatures and the conclusion the system set from the results; before approval only a draft is available. Every export is recorded in the audit trail. ### Can instruments and chromatography data systems send results? Yes, through a controlled interface that is registered, bound to an API key, validated and then made active. An instrument data system reports only for the instruments assigned to it, every message is logged as accepted, rejected or duplicate, and a received result goes through the same calibration, standard, late entry and OOS checks as a typed one. ### Is the audit trail Part 11 ready? The LIMS is built to 21 CFR Part 11 and EU GMP Annex 11: each result is attributable to the signed-in analyst, reviewers and approvers sign with password re-entry, nothing is deleted, and every change is kept in a hash-chained audit trail. The site validates it for its intended use. ### How is it licensed? Per user per month, on its own or with the electronic lab notebook and other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When th…](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Other modules - [Electronic lab notebookAn ELN for the QC laboratory: effective methods, equipment checked across LIMS, calibration and CMMS, observations never edited, results posted to the LIMS.](https://qlogix.ai/products/electronic-lab-notebook) - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) [All 22 product pages](https://qlogix.ai/products) --- # GMP electronic lab notebook connected to the LIMS | QLogix Source: https://qlogix.ai/products/electronic-lab-notebook # Electronic lab notebook (ELN) for the GMP QC laboratory The QLogix electronic lab notebook (ELN) is the analyst's contemporaneous record of laboratory work, kept on the same audited, signed records as the LIMS and the quality system. Each entry records an analytical run, a solution preparation, a method verification or investigation testing: the effective method from document control, the LIMS samples and tests, every piece of equipment checked for fitness across the LIMS, calibration management and the CMMS, released and in-date reference standards, reagents with lots and expiry, the procedure steps, time-stamped observations that are struck through with a reason but never edited, system suitability, and the results, posted straight to the LIMS samples where the specification judges them. The analyst completes the run; a second person reviews and signs it, and the entry is locked. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Five kinds of entry: analytical run, solution or reagent preparation, method verification or transfer, investigation testing, and general entry, each numbered ELN-year-sequence with its analyst, laboratory, purpose and start time. - The method chosen from the effective controlled documents in document management, with the version in force; a method outside the QMS (a pharmacopoeial chapter, a supplier method) is recorded as a reference. - The LIMS samples and the tests the run performs on them, linked both ways, so the sample shows its notebook entries. - Equipment checked for fitness when it is added and again when the run is completed: the LIMS instrument's status and calibration, the calibration record's due state and open out-of-tolerance results, the maintenance asset's status and open work orders. Equipment that is not fit is refused. - Reference standards from the LIMS, refused unless released and in date; reagents and solutions with grade, supplier, lot and expiry, and how each prepared solution was made. - Procedure steps marked done or not applicable with a reason; observations, calculations and deviations from the method time-stamped by the server, never edited, struck through by their author with a reason. - System suitability criteria with their acceptance and results; determinations entered against the sample's tests and posted to the LIMS, where the specification judges them and an out-of-specification result is taken to an investigation. - Raw data (chromatograms, spectra, printouts, worksheets) attached to the entry and fingerprinted; completion blocked until every step, criterion, result and conclusion is in place; second-person review and approval with electronic signatures. ## How it fits the platform The ELN sits between document control, the LIMS and the equipment records. The method comes from document management and only its effective version can be followed. Results go to the LIMS sample, which keeps them with the entry's reference, so the reviewer of the sample opens the run behind each result and its raw data. One piece of equipment is known to the LIMS as an instrument, to calibration management as a calibration record and to the CMMS as an asset; the ELN reads all three, and the LIMS instrument can be linked to its calibration record and maintenance asset so the LIMS refuses a result on it while either says it is not fit. Every connection shows from both sides: the sample lists its notebook entries, the instrument, calibration record and maintenance asset list the runs that used them, a failed or out-of-tolerance calibration lists the runs to assess for impact, and the controlled document lists the runs that followed it. ## Where people decide The analyst records the work as it is done and completes the run; the system checks the equipment, the standards and the reagents, judges the results against the specification and refuses what does not comply. A reviewer who did not perform the run checks the record against the raw data and signs; QA approves the sample results in the LIMS. ## Where people decide - Equipment, standards and reagents that are not fit for use, out of date or expired are refused when they are added, and equipment is checked again at completion. - Observations are never edited: their author strikes them through with a reason, and the original stays legible. - The run is reviewed and signed by a person who did not perform it; the approval locks the entry. ## Built to - 21 CFR 211.194 laboratory records - 21 CFR 211.160 and 211.165 laboratory controls - 21 CFR Part 11 and EU GMP Annex 11 - EU GMP Chapter 6 Quality Control - MHRA GxP data integrity guidance - ALCOA+ data integrity principles ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Electronic lab notebook ### Does the ELN replace the LIMS? No. The LIMS holds the samples, specifications and results that QA approves; the ELN holds how each result was obtained. Results are entered once, in the notebook, and posted to the LIMS sample, where the specification judges them. ### What happens when equipment is out of calibration? It is refused when the analyst adds it to the run, with the reason from the module that stops it: an overdue or failed calibration, an undecided out-of-tolerance result, an asset under maintenance or out of service. If it fails during the run, the check at completion shows it and the analyst records a deviation from the method. ### Can an observation be corrected? It is never edited. Its author strikes it through with a reason and records a new one; both stay on the record with their times and names, and the audit trail keeps each step. ### How is it licensed? Per user per month, on its own or with the LIMS and other modules, ordered by purchase order from the client portal. ## Reading - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When th…](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Other modules - [MES (electronic batch records)MES for pharma with electronic batch records: your master batch record, every step signed and verified, review by exception, ERP orders and QA release.](https://qlogix.ai/products/mes-electronic-batch-records) - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) [All 22 product pages](https://qlogix.ai/products) --- # MES for pharma with electronic batch records (EBR) | QLogix Source: https://qlogix.ai/products/mes-electronic-batch-records # MES for pharma: electronic batch records signed step by step The QLogix MES is a manufacturing execution system that runs electronic batch records from the client's own master batch record, for therapeutic products, medical devices and combination products. The AI drafts the master from the existing batch record, the engineer edits it and QA approves it with signatures. Every step of an executed batch record is executed by an operator who enters the required fields and signs with their credentials, and verified by a second person who signs; an entry outside its limits is recorded as entered and flagged, and needs a deviation with its investigation. QC samples go to the LIMS, orders and material consumption are exchanged with the ERP, and QA reviews the batch by exception and signs the release. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Master batch records per product: code, kind, presentation, standard batch size, version, site and the bill of materials, with steps in phases from weigh and dispense through compounding, transfer to fill-finish, fill-finish, packaging and labelling, or the phases of the client's own outline. - AI drafting from the client's batch record outline (the paper batch record or an export, the procedure, the bill of materials, specifications, artwork) in the client's wording; it never invents a limit or a code the outline does not give, and returns questions for what is missing. - Step entries: numbers with units and limits, lot numbers, equipment IDs, yes/no checks with the expected answer, choices, dates and times, attached printouts; critical steps always verified by a second person. - Execution with credential signatures: the operator saves entries as the work is done and signs Executed with email and password; the signature is refused while a required entry is missing, an earlier step is incomplete, the batch is on hold, or a lot is not among the released, unexpired lots the ERP sent. - Verification by a second person who signs Verified or returns the step with what must change; every return stays on the record, and a step marked not applicable carries a signed reason. A critical step is never skipped. - Entries outside their limits flagged at once and never corrected to pass; deviations raised on the batch, tied to the step and naming the investigation that handles them. - Review by exception: the record lists every exception (entries outside limits, deviations, returned steps, steps not applicable, holds, comments on signed steps, QC samples out of specification) and QA reviews each one, with a written conclusion on those of high severity. - ERP integration through the API: manufacturing orders in, the production schedule and material consumption per batch and step out, and the released, quarantined, rejected or consumed lots that dispensing accepts. - Batch release: disposition as approved for release, approved with conditions or rejected; review and release with electronic signatures; the executed batch record exported as a locked PDF, every export logged. ## How it fits the platform A deviation on a batch names its investigation in the quality system, and the batch cannot be released while a flagged entry has no deviation or a deviation has no investigation; release without conditions waits until every investigation the batch raised is closed. In-process and release samples are registered in the LIMS against the specification in force for the product at that stage, linked both ways, and release testing must be approved in the LIMS before the disposition. With an API key carrying the MES scope, the ERP creates batches from manufacturing orders, reads the schedule and the consumption, and sends the lots on file, so the lot an operator enters for a material is checked against what the ERP has released. A change to the process is a new version of the master; batches already started keep the version they were made from. ## Where people decide The AI drafts the master batch record; the engineer edits every phase, step and entry, and QA approves it with signatures. Operators execute and sign each step, a second person verifies it, and the supervisor can hold and resume the batch. QA reviews each exception, records the disposition with a reason and signs the release; the system flags and refuses, but never dispositions a batch. ## Where people decide - A master batch record is in force only once QA approves it with signatures; a process change is a new version, and batches already started keep theirs. - The person who executed a step cannot verify it, and nobody who executed or verified a step reviews its exceptions or approves the release. - An entry outside its limits is recorded as entered and flagged, never corrected to pass; the batch waits for a deviation and an investigation on it. ## Built to - 21 CFR 211.186 and 211.188 master and batch production records - 21 CFR 211.192 production record review - EU GMP Chapter 4 Documentation - 21 CFR Part 820 and ISO 13485 for device production records - 21 CFR Part 11 and EU GMP Annex 11 ## White paper Lab and plant, connected LIMS, calibration, logbooks and CMMS on one AI-enabled infrastructure [Download the white paper](https://qlogix.ai/whitepapers#connected-lab-and-plant) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about MES (electronic batch records) ### What is an electronic batch record? An electronic batch record (eBR) is the executed record of one lot, made against an approved master batch record and captured electronically with signatures instead of on paper. In the QLogix MES each step is executed and signed by an operator, verified and signed by a second person, and the completed record is reviewed and released by QA with electronic signatures. ### Do we have to rebuild our master batch record? No. Attach the existing paper batch record or an export from another system, with the procedure and the bill of materials, and the AI drafts the master step by step in your wording. It never invents a limit or a code the outline does not give; the engineer edits the draft and QA approves it. ### What does review by exception mean in the MES? The batch record lists every exception: entries outside their limits, deviations, returned steps, steps marked not applicable, holds, comments on signed steps and QC samples out of specification. QA reviews each one, with a written conclusion on the high-severity ones, while the steps executed within limits and verified without a return are accepted on the system's controls. The reviewer cannot be anyone who signed a step on the batch. ### How does the MES connect to our ERP? Through the REST API with a key carrying the MES scope. The ERP sends manufacturing orders and the status of material lots, and reads the production schedule and the material consumption per batch and step. Once lots are on file for a material, dispensing accepts only its released, unexpired lots. ### Is the audit trail Part 11 ready? Every signature on a step is an electronic signature under 21 CFR Part 11: the signer types their sign-in email and password each time, and the signature records the printed name, date and time, meaning and method. The audit trail keeps every entry saved, every signature and every failed signature attempt, and five wrong passwords lock signing for 15 minutes. ### How is it licensed? Per user per month, on its own or with the LIMS and other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing mode…](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) - [AI in deviation investigations: faster, more consistent root cause analysisAI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable …](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) ## Other modules - [ERPA GMP ERP on audited records: QC on receipt, lots released by QA with genealogy, bills of materials as unit operations, work orders, scheduling and planning.](https://qlogix.ai/products/erp-materials-inventory) - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) [All 22 product pages](https://qlogix.ai/products) --- # GMP ERP: materials, inventory and work orders | QLogix Source: https://qlogix.ai/products/erp-materials-inventory # ERP for GMP manufacturing: materials, inventory, bills of materials and work orders The QLogix ERP runs the material and product flow of a GMP plant on the same audited, signed records as the quality system, the LIMS and the MES. The item master carries each material's and product's unit, shelf life, storage, QC on receipt against the LIMS specification, stock policy, standard cost and approved suppliers. Purchase orders are received into lots that start in quarantine, with the LIMS sample registered at once; QA releases, rejects or holds each lot with an electronic signature, and the quantity on hand is computed from the lot's movements, never typed. Bills of materials are versioned and approved, their stages written as unit operations with lines, outputs, times, resources, QC tests and yield limits. Work orders follow the product stage by stage from dispensed materials through intermediates and bulk to the finished lot, with the schedule, the QC windows, the resource load, the budget (direct costs computed, overhead allocated by the AI with its arithmetic shown) and the MES batch record. Sales orders ship from released lots, and the material plan lists what to buy or make, read by the AI and ordered by a person. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - An item master and approved suppliers: materials (API, excipients, solvents, packaging components, labels, consumables) and products (intermediate, bulk, finished) with unit, shelf life and retest period, storage, hazard, QC on receipt against a LIMS specification, minimum stock, reorder quantity, lead time, standard cost, and up to 20 suppliers with their qualification documents; a GMP-critical item is ordered only from an approved supplier. - Purchasing and QC on receipt: purchase orders sent to the supplier and received line by line; each receipt is a new lot in quarantine with its expiry from the shelf life, its LIMS sample registered against the item's specification, and an over-delivery accepted only on purpose. - Inventory and genealogy: every lot with its origin, supplier lot, location, QC sample, decisions and movements; the quantity on hand computed from receipts, production, returns, issues, consumption, shipments, scrap and adjustments; the state read from the status and the dates (awaiting QC, quarantine, released, on hold, rejected, expired, retest due, consumed); and the parents of every in-process and finished lot, so a finished lot traces to every material lot it consumed. - QA's signed release: a lot is released, rejected or held with the signer's credentials; the release waits for the LIMS sample's approval unless QA overrides with a reason that is recorded; a rejected lot is never released again; a released lot is mirrored to the MES, so dispensing on a batch record accepts only ERP-released, unexpired lots. - Bills of materials as unit operations: templates by product category (oral solid dose, sterile injectable, lyophilized, biologic drug substance, cell or gene therapy, medical device and more) and product BOMs that start from them; stages in order (dispensing, granulation, blending, compounding, filtration, filling, lyophilization, inspection, packaging, labelling) with their lines, output, yield limits, setup, run and hold times, the resources they book and the QC tests they trigger; costing with labour, machine and QC rates and the site's overhead pools; an AI draft from the existing bill of materials, master formula or batch record that never invents a quantity; an AI review before QA approves; versions that supersede on approval. - Work orders from dispensing to finished goods: requirements computed from the approved BOM for the quantity, scrap included; materials issued from released lots only; stages started and completed in order, each producing its in-process lot with the consumed lots as parents; the yield checked against the BOM's limits and, when outside, blocked until a deviation is linked or QA accepts it with a signed reason; the finished lot in quarantine for QA's release; reconciliation and closure. - Scheduling and resources: the plan stage by stage from the planned start with setup, run and hold times, each QC test planned and due after the laboratory's turnaround, the release-ready date, resource bookings, conflicts between work orders on the same line, room or equipment, and the load per week against each resource's capacity. - Budgeting with AI overhead: direct material, labour, machine and QC costs computed by the system from standard costs and rates, every line with its basis; overhead allocated by the AI from the BOM's pools (rate = annual amount ÷ annual basis volume, amount = rate × the order's basis), the arithmetic shown and the assumptions listed; whatever cannot be priced listed, never guessed. - Sales and shipments: sales orders by customer with their lines, shipped only from released, unexpired finished lots, each shipment a movement on the lot with carrier and tracking. - Planning: per item, the released, quarantined and on-hold stock, what open work orders reserve and still need, what is on order, the projected position against the minimum stock and the shortage to buy or make; the AI writes the priorities and the risks, and a person creates the orders. - Seven alert rules with thresholds the organization sets: lots expiring or expired, lots waiting in quarantine, released stock below the minimum, work orders short of materials, yields outside limits, resources booked twice, purchase orders past their expected date. ## How it fits the platform The ERP and the MES are one record graph. A work order opens the batch record on the product's master batch record in force, the batch accepts only the lots the ERP has released, and the executed material entries are read back as consumption on the lots. Every QC sample the ERP needs, on receipt, in process or at release, is registered in the LIMS against the approved specification and followed on the lot or the work order; the release waits for the LIMS approval. A yield outside its limits or a rejected lot is investigated as a deviation in the quality system, and a new version of a bill of materials carries its change control. A resource is a CMMS asset, so its maintenance and qualification state sits beside its bookings. Q Intelligence lists and reads items, bills of materials, lots, work orders and orders alongside every other record, and the same three tools the MCP server exposes (inventory status, material requirements, work order status) are available to connected systems through the REST API. ## Where people decide The system computes quantities, requirements, schedules, yields and direct costs and refuses what the rules forbid; the AI drafts, reviews, allocates and explains; people decide. QA releases every lot and accepts every yield outside its limits with an electronic signature, QA approves every bill of materials and its versions, and the planner creates every purchase order and work order the plan calls for. No quantity on hand is ever typed, and no release, acceptance or approval happens without a named person, a reason and a signature. ## Where people decide - QA releases, rejects or holds every inventory lot with an electronic signature; a release before the LIMS sample is approved needs an override reason that is recorded with the decision. - QA accepts a yield outside the bill of materials' limits with a signed reason, or the work order waits for the deviation that investigates it. - QA approves each bill of materials and each new version before any work order runs on it; the AI's draft and review are inputs, never the approval. - Planners create the purchase orders and work orders the plan suggests; the AI reads the plan and writes the priorities but orders nothing. ## Built to - 21 CFR 211.80 to 211.94 control of components and containers - 21 CFR 211.110 in-process controls and 211.165 release testing - 21 CFR 211.184 component records and 211.186 master production records - EU GMP Chapter 4 Documentation and Chapter 5 Production - ICH Q7 for active pharmaceutical ingredients - ICH Q9(R1) Quality risk management - 21 CFR Part 11 and EU GMP Annex 11 ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about ERP ### What makes this a GMP ERP rather than a general ERP? Every lot is released by QA with an electronic signature after its LIMS sample is approved, only released and unexpired lots can be issued or shipped, the quantity on hand is computed from audited movements rather than typed, every in-process and finished lot carries its parents, and every change, decision and signature is in a hash-chained audit trail. Finance is left to the finance system. ### How does the ERP connect to the MES and the LIMS? A work order opens the batch record on the product's master batch record in force and reads the batch's executed material entries back as consumption. Every released lot is mirrored to the MES, so dispensing accepts only ERP-released lots. QC on receipt, in-process and release samples are registered in the LIMS against the approved specification, and a lot's release waits for the sample's approval. ### What does the AI do, and what does it never do? It drafts a bill of materials from the existing bill of materials, master formula or batch record without inventing a quantity, time or test; it reviews the bill of materials against the process and GMP expectations before QA approves it; it allocates overhead to a work order's budget with the arithmetic shown; and it reads the material plan and writes the priorities. It never releases a lot, accepts a yield, approves a bill of materials, creates an order or changes a quantity. ### Does it replace our finance system? No. There is no general ledger, payables, receivables or invoicing, and the budget is a planned cost at standard rates, not a costing of actuals. Purchase and sales orders carry prices and documents; the ledger stays where it is. ## Reading - [AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing mode…](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) - [EU GMP Annex 22: what the draft AI annex asks of GMP manufacturersDraft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined i…](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft) ## Other modules - [GMP logbooksElectronic logbooks for GMP cleanrooms, equipment and instruments: system-stamped entries, second-person verification, never edited, periodic reviews signed.](https://qlogix.ai/products/gmp-electronic-logbooks) - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) [All 22 product pages](https://qlogix.ai/products) --- # Electronic logbooks for GMP equipment and cleanrooms | QLogix Source: https://qlogix.ai/products/gmp-electronic-logbooks # Electronic logbooks for GMP cleanrooms, equipment and instruments QLogix GMP logbooks are electronic logbooks for cleanroom cleaning, equipment use and QC instruments, kept as audited records on the same platform as the quality system. Each entry is made by the person who did the work, stamped with the system time and verified by a second person where the procedure requires it. Entries are never edited: a wrong entry is voided with a reason and stays visible. Every book names the procedure that governs it and the review cycle that procedure sets, daily, weekly or monthly, and each period is reviewed, concluded and approved with electronic signatures. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Three kinds of book: the cleanroom cleaning log, the equipment logbook and the QC instrument logbook, each opened with its identifier, location, governing procedure, review cycle and review deadline. - Cleanroom cleaning logs: routine cleaning, disinfection with the agent class so rotation shows, periodic deep cleans, cleaning after an intervention, spill cleans, room in use and status changes, counted against the cleaning programme the procedure requires. - Equipment logbooks: use with product, batch, operation, start and end; minor, major and periodic cleaning; set-up and changeover with line clearance; maintenance, calibration, breakdowns and status changes, with clean and dirty hold times as settings. - Instrument logbooks: analyses with sample IDs, method, analyst and data reference; calibration and performance checks with their result; maintenance, service visits, malfunctions and status changes. - Entries that record who made them, when the activity happened and when the system took them; an entry later than the book's late-entry limit needs a reason and is flagged at review; second-person verification is never by the author. - A status board showing which rooms, equipment and instruments are released for use, set by the status-bearing entries; a use recorded while the subject is not released is a major finding. - Periodic review: the review falls due when the period ends, the period's entries are frozen with a SHA-256 fingerprint, and the system checks late and voided entries, verification, time sequence, failed checks, hold times and the cleaning programme. - The AI reads the period against the governing procedure and returns findings with the entries they rest on and the questions to ask the area before concluding. - Review outcomes of no issues, observations noted or deviation required, approved with signatures by an author, reviewer and approver who are different people; overdue reviews raised and escalated in the alert bell. ## How it fits the platform Logbooks sit beside the LIMS, calibration management and the CMMS on the same audited records and the same hash-chained audit trail. When a review finds a failure that may have affected product or data during GxP operation, the deviation investigation opens from the review in the quality system and links back to it. Review deadlines are watched by the alert rules for the logbooks area, so a review that is due is raised when the period ends and escalated to critical once the deadline passes. A book that is full, or whose room, equipment or instrument is decommissioned, is closed, and its last period can be reviewed at once. Where earlier periods were not reviewed on time, one review can cover several consecutive periods; the review says so, and that is itself a finding. ## Where people decide The person who did the work records it, and a second person verifies the cleanings and checks the procedure names. The system runs its GMP checks and the AI reads the period, but the reviewer concludes, and a major finding cannot be concluded as no issues without being addressed. The approver's signature fixes the book as reviewed through the period. ## Where people decide - Entries are never edited: a wrong entry is voided with the reason, stays visible struck through, and is replaced by a new entry. - Cleanings and checks the procedure wants confirmed are verified by someone other than the author. - A major finding cannot be concluded as no issues without being addressed, and the review's author, reviewer and approver are different people. ## Built to - 21 CFR 211.67 and 211.182 equipment cleaning and use logs - EU GMP Chapter 4 Documentation (logbooks) - EU GMP Annex 1 cleaning and disinfection - 21 CFR Part 11 and EU GMP Annex 11 - MHRA GxP data integrity guidance - ALCOA+ data integrity principles ## White paper Lab and plant, connected LIMS, calibration, logbooks and CMMS on one AI-enabled infrastructure [Download the white paper](https://qlogix.ai/whitepapers#connected-lab-and-plant) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about GMP logbooks ### What is an electronic logbook in GMP? An electronic logbook is the GMP record of what happened in a room or on a piece of equipment or an instrument (cleaning, use, checks, maintenance, malfunctions and status changes), kept as an audited electronic record instead of a bound paper book. In QLogix each entry is attributed to its author, stamped by the system, verified where the procedure requires it, and reviewed periodically with signatures. ### Can a logbook entry be corrected? It is never edited. A wrong entry is voided with the reason, stays visible struck through, and a new entry replaces it; the audit trail keeps each step with the name and time. ### How often are logbooks reviewed? On the cycle the governing procedure sets: daily, weekly (Monday to Sunday) or monthly, with a number of days after the period end by which the review must be approved. The book shows the review as due and then overdue, and the alert bell escalates it once the deadline passes. ### What does the system check at periodic review? Late and voided entries, entries without second-person verification or verified by their author, entries out of time sequence, uses without an end time, failed checks, and use while the subject was not released. It also checks each cleanroom programme task against its entries, equipment clean and dirty hold times, and instrument analyses after a failed check without a passing one. ### Is the audit trail Part 11 ready? Logbooks are built to 21 CFR Part 11 and EU GMP Annex 11: entries are attributable and system-stamped, never deleted but voided with a reason, the reviewed entries are frozen with a SHA-256 fingerprint, and reviews are approved with electronic signatures. Every change is kept in the platform's hash-chained audit trail. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, avai…](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [AI in deviation investigations: faster, more consistent root cause analysisAI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable …](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) ## Other modules - [PM LibraryA particulate matter library for fill-finish: defects characterized with images and FTIR, approved by QA, and current kits for inspector and AVI qualification.](https://qlogix.ai/products/pm-library) - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) [All 22 product pages](https://qlogix.ai/products) --- # Particulate matter library for visual inspection | QLogix Source: https://qlogix.ai/products/pm-library # Particulate matter library for visual inspection in parenteral fill-finish The QLogix PM Library is a particulate matter and defect library for parenteral fill-finish: every particle and defect type the filling process has produced, characterized with high-resolution images and its material identity from FTIR, Raman or SEM-EDX, placed where it enters the product relative to the 0.2 µm sterilizing filtration, and approved by QA. Approved defects form defect sets for each product, container and line, used as the kit that trains and qualifies manual inspectors and as the challenge set for automated visual inspection (AVI). Training, qualification and AVI challenge sessions record the detections per defect, judge each participant against the procedure's criteria and are approved by QA with signatures. A re-characterization is a new version, so the sets that hold the old one show as outdated and the next requalification runs against the current process. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Library entries by class: extrinsic, intrinsic and inherent particulates, and container, cosmetic or fill defects, with criticality, visibility, morphology, color and size range; inspection artefacts such as bubbles can be recorded so inspectors learn what is not a defect. - Evidence on each entry: high-resolution PNG, JPEG or TIFF images up to 20 MB each, the FTIR spectrum, Raman or SEM-EDX reports, and the material identity with method, instrument, characteristic bands, their assignments and the library match. - AI characterization from the images and the analytical data: a proposed class and identity with the rationale, the introduction point and likely source, the inspection cue, the controls and the questions; the analyst records what is confirmed. - Introduction points mapped relative to the 0.2 µm sterilizing filter: upstream bulk and compounding, the filter assembly, filling equipment, primary packaging components, environment and personnel, stoppering, lyophilization and capping, product-inherent formation after filling, and downstream handling. - The inspection cue on each defect: what the inspector looks for, how reliably manual inspection and AVI detect it, and the reject category. - QA approval with an electronic signature, after which the entry is locked; a re-characterization is a new version that supersedes the old one, and a retired defect leaves with its reason. - Defect sets per product, container, line and site, with the procedure, the effective day and the basis, holding approved defects with their units and their role in training, testing or both; each set needs at least three defects and at least one critical defect. - An AI coverage review of each set against the approved library and the linked deviations: missing critical and post-filtration defects, classes and introduction points without an example, quantities too small to measure detection, superseded or retired members. - Inspector training, qualification and requalification, AVI challenge tests and AVI recipe verification against an approved set, with detections per defect and false rejects judged against the procedure's criteria (by default 100% critical, 90% major and 70% minor detection, at most 5% false rejects). ## How it fits the platform A defect first seen in a deviation or a complaint is characterized from that record, and the library entry links back. Knowing which side of the filter a particle entered on directs the investigation and the CAPA: a particle the filter would retain, found in a filled unit, entered after the filter, from the filling equipment, the components, the environment or the closure operations. Deviations, investigations, CAPAs and change controls, such as a new component supplier, link to the defects they produced, and Q Intelligence answers questions such as which defects a kit lacks. Approved sessions are the qualification record of each inspector and each AVI recipe against the defect set in force at the time. When the library changes, the set is revised as a new version, and the next requalification runs against the updated process. ## Where people decide The AI proposes the characterization and points at the gaps in a set; the analyst records the identity, the introduction point and the source, each value with a reason once set. QA approves every library entry, every defect set and every session outcome with an electronic signature, and the session records the actions for anyone who failed: retraining, restriction from inspection or an AVI recipe change. ## Where people decide - The AI proposes a defect's class, identity and introduction point; the analyst records what is confirmed, and QA approves the entry before it can join a set. - A defect set is routed for QA approval only with at least three approved defects, at least one of them critical, each with its units in the kit. - Each participant's pass or fail is computed from the detections against the procedure's criteria; QA approves the session and the actions for anyone who failed. ## Built to - USP <1>, <790> and <1790> - FDA draft guidance: Inspection of Injectable Products for Visible Particulates (December 2021) - EU GMP Annex 1 (2022) section 8, inspection of filled containers - 21 CFR 211.160(b) and 211.110(c) and (d) - 21 CFR Part 11 and EU GMP Annex 11 ## White paper From particle to qualified inspection An AI-enabled particulate matter library for fill-finish [Download the white paper](https://qlogix.ai/whitepapers#pm-library-fill-finish) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about PM Library ### What is a particulate matter library? A particulate matter library is a controlled collection of the particles and defects a fill-finish process has produced, with images, material identity and the point where each enters the product, used to train and test visual inspection. In QLogix each entry is characterized, approved by QA, versioned rather than overwritten, and drawn into the defect sets used for inspector qualification and AVI challenge tests. ### How do you qualify visual inspectors? Against an approved defect set for the product, container and line. A qualification or requalification session presents each inspector with the set's defects in the kit quantities and acceptable units, records the detections per defect and the false rejects, and judges the outcome against the procedure's criteria. QA approves the session, with the actions for anyone who failed. ### Can the same library challenge an AVI machine? Yes. A defect set can serve as the inspector kit, the AVI challenge set or both, and AVI challenge tests and recipe verifications use the same session structure as inspector qualification. EU GMP Annex 1 expects automated inspection to detect known defects at least as well as manual inspection, and one library keeps both measured against the same defects. ### How does the library help a particulate investigation? Each approved defect records its material identity and its introduction point relative to the 0.2 µm sterilizing filter, with the likely source and the evidence. A particle the filter would retain, found in a filled unit, entered after filtration. FTIR identifies organic materials; metals and glass need SEM-EDX. ### What happens when the process changes? A new defect is characterized and approved, or an existing one is re-characterized as a new version that supersedes the old. Sets that hold an outdated or retired member are flagged as needing an update, the AI coverage review lists the gaps, and the revised set's approval means the next requalification runs against the current process. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Robotics in aseptic manufacturing: what Annex 1 expectsThe revised EU GMP Annex 1 names robotic systems, with isolators and RABS, as technologies to consider for protecting sterile products. Taking the operator out of the critical zone…](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1) - [AI in deviation investigations: faster, more consistent root cause analysisAI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable …](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) ## Other modules - [Stability Study ManagementStability studies on the LIMS: barcode pulls, reminder and escalation emails, and ICH Q1E shelf life computed in seconds by validated code, not Excel or JMP.](https://qlogix.ai/products/stability-study-management) - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) [All 22 product pages](https://qlogix.ai/products) --- # Stability study management software (ICH Q1A, Q1E) | QLogix Source: https://qlogix.ai/products/stability-study-management # Stability study management software for ICH stability programs QLogix Stability Study Management runs a pharmaceutical stability program in one system, on the same audited records as the LIMS and the QMS. Protocols are built on the approved LIMS specification, with ICH Q1A(R2) storage conditions, time points, pull windows and the tests at each time point. Studies compute the pull schedule from each batch's T0 and print a barcode for every pull; scanning the label at the qualified chamber records the pull and registers the LIMS sample with the time point's tests in one step. Reminder and escalation emails go out before each window opens and when a pull is missed. The shelf life is estimated under ICH Q1E by the system in seconds (regression, poolability at the 0.25 level, the 95% confidence bound, the extrapolation limit), with out-of-trend detection and significant change, instead of manual calculation, Excel or JMP. QA approves the protocol and the study report with electronic signatures. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Stability protocols on the approved LIMS specification: its tests become the protocol's attributes with their limits, each trended in the direction it moves (or both), with a linear or log (first-order) model, and assays flagged for the 5% significant change. - ICH Q1A(R2) and Q1B conditions as presets (25 °C/60% RH, 30 °C/65% RH, 30 °C/75% RH, intermediate, accelerated 40 °C/75% RH, refrigerated, frozen, photostability) or your own, each with time points, pull windows, units and a test matrix by time point. - An AI review of the protocol against ICH Q1A(R2), Q1B, Q1E and Q5C and the climatic zones of the markets, before QA approves it; a change is a new protocol version, and running studies keep the version they started on. - Batches placed in qualified chambers: each chamber is a CMMS asset, and placement is refused while its temperature or humidity calibration is overdue, an out-of-tolerance result is undecided or the asset is out of service. - Time-point scheduling: the pull schedule is computed from each batch's T0, by calendar month, with each pull's window, units and tests, and a Code 128 barcode on every pull label. - Barcode sample tracking: a scan at the chamber finds the pull across every study, checks its window and the chamber's state at that moment, records the time and units, and registers and receives the LIMS sample with only the tests scheduled at that time point; a pull outside its window needs the reason. - Sample pull reports: every scheduled pull across studies, missed, due now or in the next days, printed by day with the chamber, location and barcode. - Email notifications with escalation: the pull team and the study manager are reminded ahead of each window, when it opens and on its last day; a missed pull is escalated to the stability manager and QA, and raises a critical alert. - Results read back from the LIMS by condition, attribute, batch and time point, with out-of-specification results and results still in review marked; only QA-approved results enter the statistics. - ICH Q1E statistics computed by the system: regression per batch, poolability by analysis of covariance at 0.25 (slopes, then intercepts), the 95% confidence bound for the mean, the shelf life where it crosses the limit and the supported shelf life within the extrapolation limit, with a regression chart for each attribute. - Out-of-trend detection by regression control chart (the prediction interval of each batch's earlier results) and significant change at accelerated and intermediate conditions, each out-of-trend result dispositioned with an investigation or a reason. - The study report: an AI-drafted narrative that quotes the system's numbers, the shelf life and storage statement stated by the study manager, and QA approval with an electronic signature; a locked PDF with the results, the evaluation and the charts. ## Why one system for stability data Stability programs often live across a spreadsheet of pull dates, the LIMS, a statistics package and a report in a word processor. Every handoff is a transcription and a chance to miss a pull, copy a result wrong or run a regression in unvalidated software. Here the schedule, the samples, the results, the statistics and the report are the same records: a pull cannot be missed without an alert and an escalation, a result is never retyped, and the shelf-life calculation is the system's, checked against statsmodels and SciPy to 1e-6, with its inputs fingerprinted in the signed evaluation. ## How it fits the platform Pulls are LIMS samples: they are tested, judged against the specification and approved in the LIMS like any sample, results can be posted from the electronic lab notebook, and an out-of-specification result is investigated there. Chambers are CMMS assets with their calibration records, so a chamber out of calibration blocks placement and is flagged at the pull. An out-of-trend result opens an investigation pre-filled from the study, and a change control (a new supplier, a site transfer, a process change) can link to the commitment study it needs. Q Intelligence lists studies with missed pulls, open out-of-trend results or a shelf life at risk. ## Where people decide The system schedules, judges and calculates; people decide. QA approves each protocol before batches are placed on it, the analyst records each pull, a pull outside its window or from a chamber that is not fit needs a reason, each out-of-trend result needs a disposition, and QA approves the shelf life and storage statement in the study report with an electronic signature. ## Where people decide - QA approves a stability protocol (conditions, time points, tests and trended attributes) before any study is placed on it. - A pull outside its window, or from a chamber that is not fit at that moment, is recorded only with the reason, kept with the pull and in the audit trail. - Every out-of-trend result is dispositioned, with an investigation or the reason it is not out of trend, before the report can be approved. - QA approves the shelf life and storage statement the study supports, against the system's ICH Q1E evaluation. ## Built to - ICH Q1A(R2) Stability testing of new drug substances and products - ICH Q1B Photostability testing - ICH Q1E Evaluation of stability data - ICH Q5C Stability of biotechnological products - 21 CFR 211.166 Stability testing - 21 CFR Part 11 and EU GMP Annex 11 ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Stability Study Management ### What is stability study management software? Stability study management software plans and runs pharmaceutical stability programs: the protocol, the storage conditions and time points, the pull schedule, the samples and results, the shelf-life statistics and the report. In QLogix it is built on the LIMS, so every pull is a LIMS sample tested against the approved specification, and the ICH Q1E statistics are computed from the approved results. ### How is the shelf life calculated? Under ICH Q1E: each trended attribute is regressed on time per batch, poolability is tested by analysis of covariance at the 0.25 level (slopes, then intercepts), and the shelf life is where the 95% confidence bound for the mean crosses the acceptance limit, the shortest over the batches. The supported shelf life is held to the extrapolation Q1E allows: twice the long-term period covered, at most 12 months beyond it. ### Do we still need Excel or JMP for stability statistics? No. The regression, the poolability tests, the confidence bounds and the shelf life are computed by the system in seconds from the approved LIMS results, and the engine is checked against statsmodels and SciPy on known datasets. The stored evaluation records the engine version and a fingerprint of the values it read, so the signed number can be traced to its data. ### How are missed pulls avoided? The schedule is computed from each batch's T0, the pull team and study manager are emailed before each window opens, when it opens and on its last day, and a pull missed is escalated to the stability manager and QA and raises a critical alert. The pull report lists every pull due across studies with its chamber and barcode. ### How does barcode sample tracking work? Each pull has a Code 128 barcode with the study, batch, condition and month. Scanning it at the chamber finds the pull across studies, checks the window and the chamber's calibration and maintenance state, and records the pull with the LIMS sample registered and received in one step. ### How are out-of-trend results detected? A result outside the prediction interval of the line through the same batch's earlier results at that condition is out of trend (the regression control chart method). Each one is recorded for a disposition: an investigation opened from it, or the reason it is not out of trend. ### How is it licensed? Per user per month, on its own or with the LIMS and other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Other modules - [Calibration managementEvery instrument, tool and gauge: an approved requirement, readings judged by the system and reviewed by a second person, out-of-tolerance impact signed by QA.](https://qlogix.ai/products/calibration-management) - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) [All 22 product pages](https://qlogix.ai/products) --- # Calibration management software for GMP sites | QLogix Source: https://qlogix.ai/products/calibration-management # Calibration management software for instruments, tools and gauges QLogix calibration management keeps one audited record per instrument, equipment, tool, gauge, test equipment, sensor or reference standard under calibration. The calibration requirement, drafted by the AI from the manual and previous certificates, is approved with electronic signatures before the asset is relied on. Each calibration is recorded with as-found and as-left readings judged against the tolerances as they are entered, reviewed and signed by a second person, and kept with its certificate; an out-of-tolerance reading opens the impact assessment and QA signs the decision. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - An asset record with identification, criticality (GMP critical, GMP, Non-GMP), location, responsible person, description and use, and the maintenance asset it belongs to. - A calibration requirement per asset: interval and its basis, procedure or method, test points with nominal values and tolerances over the range of use, traceability of the standards, provider, acceptance criteria, the out-of-tolerance rule and the rationale; AI-drafted, edited by the metrologist, approved in review and approval. - Calibrations recorded with who, when, procedure, standards used, environment, the as-found and as-left reading at every test point judged by the system, the result, the certificate number and file, the work order, and the next due date kept by the system. - Second-person review of each calibration, signed with the reviewer's password; only a reviewed calibration counts for the asset's due date and service. - Out-of-tolerance handling: the impact assessment drafted by the AI from the readings, the earlier calibrations and the linked records; QA's decision signed; the investigation linked. - Service status with a reasoned history, retirement as the final approval, locked PDF exports, and a dashboard with compliance, the state of every calibration cycle and what falls due each month. - Alerts for assets due, overdue or failed, calibrations awaiting review, and out-of-tolerance results without a decision. ## How it fits the platform A calibration record belongs to a CMMS asset when the site keeps one, so the instrument's calibrations and the equipment's work orders read as one file, and a calibration work order records its calibration once. An out-of-tolerance result opens an investigation in the quality system with the impact assessment attached. The LIMS locks out an instrument whose calibration has expired. The tolerances the AI proposes come from the process tolerance the reading serves, which it reads from the asset's description and the linked records. ## Where people decide The AI drafts the requirement and the impact assessment; the metrologist approves the requirement, the technician records the calibration, a second person reviews and signs it, and QA signs the out-of-tolerance decision. The asset's due date is computed by the system from reviewed calibrations and never typed. ## Where people decide - The calibration requirement is approved in review and approval before any calibration is recorded against it; changes afterwards need a reason. - Each calibration is reviewed and signed by a person other than the one who recorded it. - The out-of-tolerance decision is taken and signed by a person other than the recorder, with the impact assessment on record. ## Built to - 21 CFR 211.68 and 211.160(b)(4) - 21 CFR 820.72 - EU GMP Chapter 3 - ISO/IEC 17025 for the laboratories used - ISO 10012 measurement management - 21 CFR Part 11 and EU GMP Annex 11 ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Calibration management ### How are tolerances judged? As the readings are entered, against the point's tolerance written as ±x, ±x % or +a/-b, in the point's unit or a convertible one. An as-found reading out of tolerance can only be recorded as a pass after adjustment or a fail; an as-left reading out of tolerance as a fail or a pass with limitation. ### What happens when a reading is out of tolerance? The calibration appears under out-of-tolerance results with the decision pending. The AI drafts the impact assessment for the period since the last good calibration; the assessor edits it, QA signs the decision with a reason, and an investigation is linked where a result, a batch or a validated state may be affected. ### Can calibrations come from an external laboratory's certificate? Yes. The certificate is attached to the record and linked to the calibration; the readings are entered from it and judged by the system; a second person reviews them against the certificate and signs. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. ## Reading - [Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, avai…](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When th…](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Other modules - [CMMSA GMP CMMS on audited records: assets with preventive plans, work orders verified for return to service, and reliability and predicted failures from the history.](https://qlogix.ai/products/cmms) - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) [All 22 product pages](https://qlogix.ai/products) --- # GMP CMMS with reliability and predictive maintenance | QLogix Source: https://qlogix.ai/products/cmms # CMMS for GMP maintenance: assets, preventive plans, work orders and reliability The QLogix CMMS is a computerized maintenance management system built on the same audited, signed records as the quality system. Every asset carries its criticality, GxP impact, qualified state and preventive maintenance plan; every work order is requested, planned, assigned, executed with its tasks, labour, parts, downtime and findings, assessed for GxP impact and verified for return to service with electronic signatures. From that history the module computes each asset's reliability and lifecycle stage and predicts when it will fail next. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - An asset register with tag, type, location, system, criticality, GxP impact, qualification reference and requalification date, commissioning date, operating hours and run-hour counter. - Preventive maintenance plans per asset, drafted by the AI from the manual and the maintenance procedure, with tasks, intervals and the rationale; each task raises its work order when due. - Work orders of six types: preventive, corrective, calibration, inspection, emergency and modification, with priority and target response, assignment, tasks with acceptance criteria, labour, parts, downtime and findings. - Completion verification: the GxP impact assessed, requalification or a deviation raised where needed, and the verifier's signature on the return to service. - A maintenance dashboard: PM compliance, the state of every preventive cycle, work by type, status and priority, completed work with downtime and labour, failure codes and requalifications due. - Reliability and lifecycle per asset from the work orders: mean time between failures, mean time to repair, availability, the failure trend, a Weibull fit once there are enough failures, the predicted next failure and the preventive interval that would catch it. - Alerts for overdue work and preventive tasks, assets predicted to fail, and preventive tasks scheduled after the predicted failure; an AI advisory that drafts the plan changes and the replacement cases. - Decommissioning under change control: the close-out withdraws the asset's documents, deactivates its plan, cancels its open work orders, retires its instruments' calibration records and lists the controlled documents to obsolete, each step audited with the change control's reference. ## How it fits the platform A work order that finds a failure opens a deviation investigation with the findings already written. A modification goes through change control, and a change that affects the qualified state links to the validation project that requalifies it. The instruments and gauges mounted on an asset keep their calibration records in the calibration module, linked to the asset both ways, so an out-of-tolerance calibration counts toward the asset's reliability. Run-hour counters can be sent from a PLC, SCADA or MES through the REST API, so operating time is measured rather than estimated. Everything is one record graph: the inspector follows the link from the work order to the deviation to the change control to the requalification, and back to the asset, without a spreadsheet. ## Where people decide The AI proposes maintenance plans, classifies work and drafts the advisory; the maintenance engineer sets the plan, the technician records the work, and the verifier signs the return to service. Nothing changes an interval, closes a work order or returns an asset to service without a named person and a reason. ## Where people decide - Preventive plans are approved by the asset owner; a change to the interval of a qualified asset goes through change control. - Work is declared complete by the technician and verified for return to service by a second person, who signs. - The reliability model's predicted failures and recommended intervals are decision support; the engineer confirms before a plan changes. ## Built to - 21 CFR 211.67 and 211.68 - 21 CFR 820.70 - EU GMP Chapter 3 and Annex 15 - 21 CFR Part 11 and EU GMP Annex 11 - ISO 55000 asset management principles ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about CMMS ### Does the CMMS replace our existing maintenance system? It can, or it can run alongside it. Assets and work orders live in QLogix as audited records; a run-hour counter or a work order feed from another system comes in through the REST API with the source named in the audit trail. ### How does predictive maintenance work without sensors? From the work orders: corrective and emergency work counts as failures, their downtime as unavailability. With operating hours per day or a counter reading the figures are in operating time; with four failures the Weibull fit says whether the asset is wearing out and where the preventive interval should sit. ### Is the return to service signed? Yes. Completion verification is a review and approval cycle with electronic signatures under 21 CFR Part 11: the verifier records the outcome, the GxP impact and any requalification, and signs. ### What happens when an asset is decommissioned? It leaves service only under an approved change control. The close-out withdraws every document on the asset and on its open work orders, deactivates the preventive plan, cancels the open work orders, retires the linked calibration records and lists the controlled documents to obsolete. Nothing is deleted: the records and their audit trails stay, and verified work orders keep their documents as evidence. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, avai…](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Other modules - [ValidationComputer system validation software built on GAMP 5 and CSA: URS and risk, IQ/OQ/PQ protocols, electronic test execution, a live traceability matrix, signed reports.](https://qlogix.ai/products/computer-system-validation) - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) [All 22 product pages](https://qlogix.ai/products) --- # Computer system validation software: CSV, CSA, GAMP 5 | QLogix Source: https://qlogix.ai/products/computer-system-validation # Computer system validation software for paperless CSV and qualification The QLogix validation module is computer system validation software that runs the whole validation lifecycle on audited, signed records, for computerized systems and for equipment, facilities and utilities, processes and cleaning. A project starts with an AI-drafted validation plan and its activities; the validation documents, from the URS and risk assessment to the IQ, OQ and PQ protocols and the summary report, come from templates, are drafted by the AI and are approved with electronic signatures before execution. Protocols are executed step by step on screen, with the actual result, the evidence, the tester's name and the server time on every step. For a computer system, the GAMP 5 software category and the risk of each requirement set the deliverables and how each requirement is verified, and closure is refused until every requirement to be tested is verified. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Five validation types: equipment, facility and utilities, process, cleaning, and computer systems (CSV) with a GAMP 5 software category of 1, 3, 4 or 5, each project with its system, risk level, target date and change control reference. - A validation plan drafted by the AI with the approach, scope, deliverables, acceptance criteria, prerequisites and risks, and 10 to 30 activities across planning, pre-execution, execution, reporting and closure, each with an owner and a due date; assignees are emailed and one view tracks open, overdue and unassigned activities across projects. - Eighteen document types from templates that carry the Part 11, Annex 11 and good documentation practice elements: validation plan, URS, risk assessment, supplier assessment, functional, configuration and design specifications, DQ, IQ, OQ, PQ, PPQ and cleaning validation protocols, traceability matrix, test execution record, qualification report, validation summary report and protocol deviation report. - AI drafting of every section, with 4 to 20 test cases per protocol (steps, expected results and acceptance criteria); facts the AI cannot find are left as [placeholders] for the author. - Electronic test execution of an approved protocol: each step recorded as it is performed with the actual result, pass or fail and evidence, under the tester's name and the server time. A failed step ends the run, a deviation is raised, and the retest states its reason. - For computer systems, a system assessment (intended use, GxP impact, Part 11 and Annex 11, data criticality, hosting, supplier), a requirements register with a GAMP 5 functional risk assessment of severity, probability and detectability, and the deliverables GAMP 5 expects for the software category and risk. - Verification scaled to risk in line with computer software assurance: high-priority requirements tested with positive, negative and boundary tests and evidence, medium ones tested or covered by assessed supplier testing, low ones verified by inspection, review or supplier evidence with the rationale. - A readiness check before execution (environment and baseline, trained testers, checklist, pre-execution activities); only testers named on the project with a training reference can execute, and the execution is reviewed by people who executed none of the tests. - A live traceability matrix from each requirement to its tests and their latest execution, downloadable as CSV, and closure with the outcome Validated, Validated with conditions, or Not validated. ## How it fits the platform A validation project can be opened from the record that requires it, such as a change control, and a CMMS modification that affects an asset's qualified state links to the validation project that requalifies it. The risk assessment module carries a CSV FMEA template for the GAMP 5 functional risk assessment, and a validation project can also be planned and tracked in project management with its baseline and critical path. Every plan, protocol, execution record, report and deviation is a record with its signatures and audit trail, listed in document management's index with the rest of the quality system's documents. Open validation projects and documents appear in the work queue, and overdue validation activities raise an alert. The inspector follows the trace from requirement to test to executed step to signed report without a paper binder. ## Where people decide The AI drafts the plan, the activities, the requirements and their risk, and every document section; it does not approve, execute or sign. The validation lead assigns owners and dates, signers approve the plan and each protocol before execution, testers record every step themselves, people who ran none of the tests review the execution, and the validation summary report and closure are signed. A recorded outcome is never changed: a correction of an actual result or evidence carries a reason and stays in the audit trail. ## Where people decide - Protocols are approved with electronic signatures before execution; a report can only be created from an approved protocol. - Declaring a computer system ready for execution is refused until every readiness check is met, and the test environment is then fixed until the declaration is withdrawn with a reason. - Validation closure is refused until every requirement to be tested is verified through an approved execution record, every activity is done or cancelled and the validation summary report is signed. ## Built to - GAMP 5 (Second Edition) - FDA Computer Software Assurance guidance - 21 CFR Part 11 and EU GMP Annex 11 - EU GMP Annex 15 - 21 CFR 211.63 and 211.68 - PIC/S PI 041 data integrity guidance ## White paper Validation, digitized and on schedule Electronic validation with project management built in [Download the white paper](https://qlogix.ai/whitepapers#validation-and-project-management) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Validation ### What is computer software assurance? Computer software assurance (CSA) is FDA's risk-based approach to establishing that software used in production or the quality system is fit for its intended use; the final guidance was announced in September 2025 for medical device software, and drug manufacturers commonly apply the same reasoning through GAMP 5. Assurance effort follows process risk: features whose failure could compromise safety get rigorous scripted testing, the rest lighter activities and supplier evidence. In QLogix each requirement's risk priority sets how it is verified, and the rationale is recorded. ### What is the difference between CSV and CSA? Computer system validation (CSV) is the requirement that software used for GMP is validated for its intended use. CSA does not remove that requirement; it changes how the effort is spread, by risk. QLogix keeps the deliverables GAMP 5 expects for the software category and scales the testing of each requirement to its risk. ### Can it validate equipment, processes and cleaning as well as software? Yes. Equipment and facility and utility projects use DQ, IQ, OQ and PQ protocols, process validation a PPQ protocol and cleaning validation a cleaning validation protocol, all with the same plan, execution, deviation and report flow. The system assessment, requirements register, readiness check and traceability stages apply to computer system projects. ### Is test execution paperless? Yes. An approved protocol is executed on screen: each step's actual result, pass or fail and evidence are recorded as the tester performs it, with their name and the server time. A failed step needs a deviation before a retest, and the execution is reviewed by people who executed none of the tests. ### How is the traceability matrix kept? Live. Every requirement is traced to the tests that verify it and their latest execution, and the matrix downloads as CSV. Validation closure is refused while any requirement to be tested is not verified through an approved execution record. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. There is no card payment or self-serve plan. ## Reading - [Computer software assurance: what FDA's final guidance changes for validationFDA's final guidance on computer software assurance replaces test-everything validation with effort matched to risk. Features that could compromise safety get scripted testing; the…](https://qlogix.ai/blog/computer-software-assurance-final-guidance) - [AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writ…](https://qlogix.ai/blog/ai-document-management-validation-workflow) ## Other modules - [Project managementPrograms and projects on audited records: AI-drafted plans with work breakdown, resources, risks, baseline, critical path and earned value, signed at approval.](https://qlogix.ai/products/project-management) - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) [All 22 product pages](https://qlogix.ai/products) --- # Project management for life sciences quality programs | QLogix Source: https://qlogix.ai/products/project-management # Project and program management on audited quality records QLogix project management runs programs and projects on the same audited records as the quality system. A project carries its objective and scope, an AI-drafted plan with work packages, subtasks, milestones and dependencies, the resources and their weekly load, a risk register with responses and contingency plans, the schedule baseline, and its status: critical path, earned value and forecast finish. Plan approval and closure are signed electronically, and every task owner is reminded on the cadence the project manager sets. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Programs that group projects under one outcome, with benefits, program risks and dependencies across projects. - Projects with an AI-drafted plan: work packages, subtasks, milestones, dependencies (finish-to-start and the others), effort and owners; a timeline with the critical path and a draggable task column. - Resources with weekly capacity and allocation; overallocation shown and alerted. - A risk register per project with likelihood, impact, responses, triggers and contingency plans. - The schedule baseline set at plan approval; earned value, schedule and cost performance indices and the forecast finish from progress; red, amber and green status with the reasons. - Reminders to task owners ahead of the planned finish, on the day and while overdue, with escalation to the manager; dashboards for every project and program. ## How it fits the platform A validation project, a change control's implementation, a CAPA's actions or a CMMS modification can be planned and tracked as a project and linked to the records that carry the work. The plan and its baseline are audited like every other record, so what was promised, when, and what changed are on file for the inspector and for the steering committee alike. ## Where people decide The AI drafts the plan and the risk register; the project manager edits them, the sponsor approves the plan and signs, the owners do the work and record it, and closure is approved with the outcome and the lessons. ## Where people decide - Plan approval and closure are review and approval cycles with electronic signatures. - Dates, effort and cancellations on an approved plan need a reason, recorded in the audit trail. - The AI never sets a baseline, closes a task or approves a plan. ## Built to - ICH Q10 pharmaceutical quality system - EU GMP Annex 15 and GAMP 5 for validation projects - 21 CFR Part 11 and EU GMP Annex 11 - PMI practice for earned value and critical path ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Project management ### Is this a replacement for a general project tool? For quality, validation, maintenance and compliance programs, yes: the plan lives next to the records it delivers and is audited with them. For software or construction programs a general tool may still fit better; both can coexist. ### Does it do resource levelling? It shows each resource's weekly load against capacity and alerts on overallocation; the manager moves the work. It does not reschedule tasks on its own. ### How are reminders sent? By email to each task owner ahead of the planned finish, on the day and while overdue, on the cadence set on the project, with escalation to the manager past a threshold. The reminders are logged. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. ## Other modules - [Operational excellenceA3s, kaizen, value stream maps, 5S, standard work, SQDIP boards, bowlers and policy deployment on audited records, with OPEX projects and your own CI apps.](https://qlogix.ai/products/operational-excellence) - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) [All 22 product pages](https://qlogix.ai/products) --- # Operational excellence and lean CI software | QLogix Source: https://qlogix.ai/products/operational-excellence # Operational excellence software: continuous improvement planning and execution QLogix operational excellence runs continuous improvement on the same audited, signed records as the quality system. Each improvement tool is an application: A3 problem solving, kaizen events, value stream maps, 5S audits, standard work, leader standard work, the SQDIP daily management board, KPI bowlers, policy deployment, the cost-benefit ratio, the voice of the customer, transactional process improvement, change management, RACI, waste walks, improvement ideas, maturity assessments, skills matrices, kanban sizing and development stage gates. The system computes what each tool computes (takt, lead time and process cycle efficiency, the gap closed, audit scores, line balance, red and green months, the cost-benefit ratio and payback) and flags what the method expects next. The AI drafts a record from the team's description and data and coaches it against the method. Clients build their own applications from their existing forms, approve and version them. Every record carries an action plan with owners, due dates and evidence, and the projects that carry improvements out run in project management with an OPEX ID. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Twenty-one built-in applications in seven families: problem solving (A3, improvement ideas), lean and flow (kaizen, value stream mapping, 5S, standard work, transactional process improvement, waste walks, kanban sizing), daily management (SQDIP board, leader standard work), strategy deployment (KPI bowler, policy deployment X-matrix), customer and growth (voice of the customer and CTQ tree, development stage gates), people and change (change management, RACI, maturity assessment, skills and certification matrix) and planning and finance (cost-benefit ratio, action plan). - Calculations the system owns: the A3's gap closed and whether the gain is sustained; kaizen targets met and improvement in each metric's direction; takt time, lead time, value-added time, process cycle efficiency and the steps slower than takt on a value stream map; 5S scores by S and against the last audit; operators needed and line balance on standard work; hit rates and three-misses-in-a-row escalation on the SQDIP board; red and green months on the bowler; coverage gaps in policy deployment; the cost-benefit ratio, the risk-adjusted ratio, payback and net value; kanban cards; first-pass yield and the value-added share of an office process. - An action plan on every record: actions from any table row (a countermeasure, an open item, a red tag, a kaizen burst) or added directly, with an owner from the organization, a due date, a status and the evidence that closes it; overdue actions across every record on one page. - Your own applications: start from a built-in or from scratch, or attach your current form and let the AI lay out the sections, fields, tables and phases in your wording; fields of eleven kinds including scores, choices, people, dates and references between tables; totals and averages in the results; approval before use and new versions that supersede on approval, with each application numbering its records with its own prefix. - AI that drafts and coaches: a record filled from the team's description and the data, reports and photos attached, without inventing a measurement; a coaching review against the method that reads the computed results and names what to strengthen. - OPEX projects and programs: when an improvement needs a planned project, it runs in project management (work breakdown, resources, risks, critical path, earned value) designated OPEX, numbered in its own OPEX-YYYY-NNNN series, linked back to the record it came from, and listed with the improvement work. ## How it fits the platform Improvement records are records like every other in QLogix: hash-chained audit trails, review and approval with electronic signatures for the close-out, attachments and links in both directions. An A3 links to the deviation or CAPA that prompted it, a daily management board's repeated misses open an A3 that links back, a cost-benefit ratio supports the OPEX project that delivers it, and a policy deployment names the OPEX projects and bowlers that carry each priority. Q Intelligence reads improvement records alongside every quality record. ## Where people decide The system computes and flags; the AI drafts and coaches; people decide. Teams own their records and their actions, and an action is done only when its evidence is recorded. A sponsor or CI lead approves the close-out with an electronic signature. A client application is approved before anyone opens a record on it, and a change to it is a new version. ## Where people decide - The sponsor or CI lead approves each record's close-out with an electronic signature once its actions are done and its results are in. - Each client application is approved before use; a change is a new version that supersedes the old one when approved. - Every action is closed by a named person with the evidence that it is done; cancelling one needs a reason. - The AI's draft and coaching are inputs; the team decides what stays on the record. ## Built to - Lean and the Toyota Production System - A3 and PDCA problem solving - Hoshin kanri policy deployment - Six Sigma DMAIC and voice of the customer - ICH Q10 Pharmaceutical quality system: continual improvement - 21 CFR Part 11 and EU GMP Annex 11 ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Operational excellence ### Which continuous improvement tools are included? A3 problem solving, kaizen events, value stream maps, 5S audits, standard work, leader standard work, the SQDIP daily management board, KPI bowlers, policy deployment (X-matrix), cost-benefit ratio, voice of the customer, transactional process improvement, change management, RACI, waste walks, improvement ideas, lean maturity assessment, skills and certification matrix, kanban sizing, development stage gates and a plain action plan: twenty-one in all, each computed by the system. ### Can we use our own forms? Yes. Build an application from scratch or from a built-in, or attach your current form or template and the AI lays out its sections, fields and tables in your wording. You approve it before anyone uses it, its records are numbered with your prefix, and a change is a new version. ### How are improvement projects managed? In project management, designated OPEX: numbered OPEX-YYYY-NNNN in a series of their own, with the work breakdown, resources, risks, critical path, baseline and earned value of any project, linked to the improvement record they came from and listed in operational excellence. Projects created in an OPEX program are OPEX too. ### What does the AI do? It drafts a record from the team's description and the documents attached, filling only what the sources support and saying what is missing, and it coaches a record against its method using the results the system computed. It also drafts a client application from an existing form. It never closes an action, approves a record or changes a computed result. ## Other modules - [Q Intelligence, Q Agent and agentsAI for pharmaceutical quality: ask across every GxP record with cited sources, draft the next record, and hand tasks to AI agents within approvals you set.](https://qlogix.ai/products/ai-for-pharma-quality) - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) [All 22 product pages](https://qlogix.ai/products) --- # AI for pharmaceutical quality with human approval | QLogix Source: https://qlogix.ai/products/ai-for-pharma-quality # AI for pharmaceutical quality: an assistant and agents that work with human approval QLogix Q Intelligence and Q Agent are AI for pharmaceutical quality: an AI quality assistant and AI agents that work on the same audited records as the quality system, while qualified people decide and sign. Q Intelligence answers questions about any record, document, audit trail or regulation, shows every source it consulted, and drafts the next investigation, CAPA, change control or document for a person to review and create. Q Agent works a goal across the quality system and proposes actions, such as logging a complaint, opening an investigation or sending an email, for a person to approve; drafts can be allowed to run on their own, but flags, emails and schedules always wait. Scheduled agents report on a cadence and propose emails and flags for approval. Nothing is signed, closed or approved by the AI, and every AI draft and action is an entry in the hash-chained audit trail with the AI tool and model named. [Request a demo](https://qlogix.ai/contact)[All products](https://qlogix.ai/products) ## What the module does - Q Intelligence searches records, attachments, audit trails and projects, then the regulatory library, preferred websites and the open web, and lists what was consulted: every record, document, guidance passage, web page and connector call it used. - Regulatory answers cited by document, code and page from the regulatory library: FDA guidance and regulations, ICH quality guidelines, EU GMP, EMA and PIC/S indexed from the official documents, with your licensed ISPE and PDA copies searchable for your users. - Draft cards for an investigation, change control, CAPA, document or record link: Review and create opens the form pre-filled, and the record exists only when a person creates it. - Analysis in the conversation: control charts, Pareto and other charts from your data, hypothesis tests, and draft FMEAs and PFMEAs built from deviations, investigations and SOPs for the team to review. - Q Agent works a goal to a visible plan, reads the records with read-only tools, and proposes actions (log a complaint, open an investigation, draft the root cause analysis, run backlog triage or signal detection, flag a record, send an email, schedule a recurring agent), each with a reason a reviewer can check. - Two levels of autonomy: ask before every action, or create drafts without asking, under which flags, emails and schedules still wait. Actions run with the approver's own user permissions, never administrator authority. - Scheduled agents for daily huddle prep, overdue CAPA chasing, APR/PQR data packs, QMS gap watch, management review packs and inspection readiness: they read records, write a report and propose emails or flags for approval. - Alerts and watches: built-in rules for records left waiting, such as a possible adverse event without a decision or a prior-approval change without a regulatory decision, plain-language watches, and Ask Q Agent to work this. - AI final review before a record is routed: the AI reads it as the reviewers will and lists must-fix, should-fix and consider findings with a suggested fix; it changes nothing. ## How it fits the platform Q Intelligence and Q Agent read the same records as every module: complaints, investigations and CAPAs, change controls and documents, validation, risk assessments and submissions, the LIMS, ELN, MES, calibration and CMMS, with their audit trails. Administrators can connect the organization's own systems as tool connectors; the chat offers only the tools a server declares read-only, and a change in a connected system goes through Q Agent, where each action waits for approval. The REST API and the MCP server let other systems and internal agents read records and start Q Agent tasks; records are read-only through the API and every action a task proposes still waits for approval. ## Where people decide The AI drafts, searches, analyses and proposes; people create the records, approve or dismiss each action, review, decide and sign. Records Q Agent creates start in review and still need a qualified person to review and sign. Text in records, files, web pages, alerts and connected systems is read as data, never as instructions: if it contains instructions, the AI says so and does not follow them. ## Where people decide - Nothing is created, changed, linked or signed by Q Intelligence: drafts become records only when a person creates them, and alert settings change only when a person confirms. - Flagging a record, sending an email and scheduling an agent always wait for a person's approval; actions run in the approver's name and are logged in the audit trail. - Q Agent never signs, closes, approves, dispositions or decides reportability; the records it creates start in review. ## Built to - Draft EU GMP Annex 22 on artificial intelligence - FDA draft guidance on AI to support regulatory decision-making for drugs and biologics - 21 CFR Part 11 and EU GMP Annex 11 - GAMP 5 (Second Edition) - FDA Computer Software Assurance guidance - ICH Q9(R1) and ICH Q10 ## White paper One quality system, one AI layer The case for an Agentic QMS [Download the white paper](https://qlogix.ai/whitepapers#agentic-qms) ## Licensing Per user per month, on its own or with other modules, ordered by purchase order from the client portal. Every module writes to the same audited records. [Ask for a quote](https://qlogix.ai/contact) ## Questions about Q Intelligence, Q Agent and agents ### How do AI agents work in a GxP quality system? In QLogix an agent is given a goal, reads the records with read-only tools, keeps a visible plan and proposes each change as an action with a reason. A person approves or dismisses each action; under the drafts setting, drafting actions run on their own while flags, emails and schedules still wait. Actions run in the approver's name, records the agent creates start in review, and nothing is signed or closed by the agent. ### Is AI allowed in GMP? Yes, within the rules for any computerized system used for GMP: it is validated for its intended use and controlled under 21 CFR Part 11 and EU GMP Annex 11. The draft EU GMP Annex 22, not final when this page was written, would keep generative AI and large language models out of critical GMP applications and allow them in non-critical uses with qualified people reviewing the output. In QLogix the AI drafts and proposes, and qualified people review, decide and sign. ### Does the AI make quality decisions? No. It drafts, searches, trends, flags and proposes. Dispositions, priorities, reporting routes, reportability, approvals and signatures stay with qualified people, and each AI draft and action is recorded in the audit trail with the AI tool and model named. ### How do I know where an answer came from? Every answer lists what was consulted: the records, documents, guidance passages, web pages and connector calls it used. Regulatory answers cite the document, code, page and official link, and AI drafts quote their evidence exactly from the write-up or the attached documents. ### What does FDA's AI credibility framework mean for a quality assistant? FDA's January 2025 draft guidance asks for a credibility assessment, scaled to model risk, for AI that supports regulatory decisions on drugs and biologics. It sets outside its scope uses that only improve operational efficiency and do not affect patient safety, drug quality or the reliability of study results; your quality system still decides how such tools are controlled. In QLogix the AI's output is a draft or a proposal that a person reviews before anything is created or signed. ### How is it licensed? Per user per month, on its own or with other modules, ordered by purchase order from the client portal. AI use is included in the licence. There is no card payment or self-serve plan. ## Reading - [FDA's seven-step credibility framework for AI, in plain termsFDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the dec…](https://qlogix.ai/blog/fda-ai-credibility-framework) - [EU GMP Annex 22: what the draft AI annex asks of GMP manufacturersDraft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined i…](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft) - [AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing mode…](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) ## Other modules - [Agentic QMSeQMS software for pharma, biotech and medical devices: change control, CAPA and controlled documents on one audited record store, AI drafts and signed decisions.](https://qlogix.ai/products/agentic-qms) - [Root cause investigationAI-assisted deviation, OOS and complaint investigations beside your existing QMS: 5 Whys and fishbone drafted, each cause tested, a robust root cause and CAPA plan.](https://qlogix.ai/products/root-cause-investigation) - [CAPA managementGMP CAPA software: CAPAs opened from investigations, complaints, audits and changes, an AI review of the plan, evidence for each action and an effectiveness check.](https://qlogix.ai/products/deviation-capa-management) - [Complaint handlingComplaint handling software for pharma and medical devices: AI triage with quoted evidence, an adverse event check, a route set by QA and the investigation started.](https://qlogix.ai/products/complaint-handling) - [Batch record reviewBatch record review software that reviews by exception: the AI lists only the exceptions with quotes, QA dispositions each one and signs the release recommendation.](https://qlogix.ai/products/batch-record-review) - [Change managementGMP change management software for document, equipment and facility changes: AI impact and ICH Q9 risk assessment, GMP requirements per item, approval and closure.](https://qlogix.ai/products/change-control) [All 22 product pages](https://qlogix.ai/products) --- # Blog: AI and quality in life sciences | QLogix Source: https://qlogix.ai/blog # AI and quality in life sciences Regulatory updates, how AI is being adopted in GxP quality systems, and what it takes to validate it. Written by the QLogix team for quality, validation and regulatory affairs leaders. - [Manufacturing science15 October 202610 min read Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.The alert limit: a single result between two and three standard deviations from the centre line is a warning, not a deviation.2 σNelson's run rules: shifts, trends, alternation, stratification and mixtures that no single-point limit sees.8Signature per signal: QA's disposition, investigate, monitor or not a signal, with the reason on record.1Read the article](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality15 October 20269 min read Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.Read the article](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality15 October 202612 min read Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.Read the article](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [AI in quality14 October 202612 min read Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.Read the article](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) - [AI in quality14 October 20267 min read AI in deviation investigations: faster, more consistent root cause analysisAI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable hypotheses, each quoting the write-up or an attached document. The investigator then tests, decides and signs. Nothing the AI writes becomes the root cause until a qualified person says so.Read the article](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis) - [AI in quality14 October 20267 min read AI-assisted CAPA: from root cause to a verified effectiveness checkA CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same cause coming back. AI helps most before the plan is approved: proposing actions that address the confirmed cause, drafting the effectiveness criterion up front, and matching the CAPA against earlier records. People approve, implement, verify and close.Read the article](https://qlogix.ai/blog/ai-capa-management-effectiveness) - [Validation14 October 20267 min read AI in document management and validation workflows for life sciencesDocument control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. The same holds for validation: plans, protocols and reports drafted from templates and requirements, with assurance effort set by risk under GAMP 5 and FDA's CSA guidance, and a person approving every version with an electronic signature.Read the article](https://qlogix.ai/blog/ai-document-management-validation-workflow) - [AI in quality12 October 20264 min read FDA's seven-step credibility framework for AI, in plain termsFDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the decision, and how serious a wrong decision would be.Read the article](https://qlogix.ai/blog/fda-ai-credibility-framework) - [Manufacturing science12 October 20264 min read AI in drug manufacturing: where FDA stands and what to build forFDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing models. Together they show what a process model has to prove.Read the article](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda) - [Medical devices12 October 20264 min read AI-enabled medical devices: what FDA's list and guidance showFDA's public list counted 1,614 AI-enabled medical devices through June 2026, about three in four of them in radiology. Two documents shape how they are built and changed: the final guidance on predetermined change control plans and the draft guidance on lifecycle management.Read the article](https://qlogix.ai/blog/ai-enabled-medical-devices-fda) - [Manufacturing science12 October 20263 min read Robotics in aseptic manufacturing: what Annex 1 expectsThe revised EU GMP Annex 1 names robotic systems, with isolators and RABS, as technologies to consider for protecting sterile products. Taking the operator out of the critical zone removes the largest contamination source and moves the quality work to design, software and data.Read the article](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1) - [Regulatory updates12 October 20264 min read EU GMP Annex 22: what the draft AI annex asks of GMP manufacturersDraft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined intended use, independent test data and ongoing monitoring.Read the article](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft) - [Validation12 October 20264 min read Computer software assurance: what FDA's final guidance changes for validationFDA's final guidance on computer software assurance replaces test-everything validation with effort matched to risk. Features that could compromise safety get scripted testing; the rest can be covered by unscripted testing and supplier evidence.Read the article](https://qlogix.ai/blog/computer-software-assurance-final-guidance) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Signal detection with control charts in GMP | QLogix Source: https://qlogix.ai/blog/spc-signal-detection-gmp-trending Manufacturing science # Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviation By QLogix · Published 15 October 2026 · 10 min read The short answer A result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification. The alert limit: a single result between two and three standard deviations from the centre line is a warning, not a deviation. 2 σ Nelson's run rules: shifts, trends, alternation, stratification and mixtures that no single-point limit sees. 8 Signature per signal: QA's disposition, investigate, monitor or not a signal, with the reason on record. 1 ## Key takeaways - FDA's process validation guidance (2011) makes trending a continuing obligation under stage 3, continued process verification, and 21 CFR 211.180(e) and EU GMP Chapter 1 require the periodic review of data for trends; ICH Q10 names process performance and product quality monitoring as a pillar of the quality system. None of them say which chart or which rules: that is the firm's decision, and the records show it was applied. - The individuals and moving range chart with sigma estimated from the average moving range, alert limits at two sigma and action limits at three, as in ASTM E2587, fits most batch and laboratory parameters: one result per lot, no rational subgroup, a specification to compare with. - Nelson's eight rules (or the four Western Electric rules) detect what single-point limits miss: nine points on one side of the centre line is a shift, six rising or falling is a drift, fifteen inside one sigma means the limits are too wide or the data is mixed. Each rule names the pattern, which is what the investigation starts from. - A signal is not a deviation. It is a question for QA: investigate, monitor or not a signal, each with a reason and a signature. Kept on one audited record with the chart, the AI's interpretation and the linked investigation, the dispositions are the trending history an inspector asks for at the annual product review. ## What do the regulations ask for in trending? For drugs, 21 CFR 211.180(e) requires written procedures for the evaluation, at least annually, of the quality standards of each drug product, with a review of a representative number of batches and of complaints, recalls, returns and investigations, to determine the need for changes in specifications or manufacturing or control procedures. EU GMP Chapter 1 asks the same of the product quality review, with an explicit review of critical in-process controls and finished product results, and 'a review of all batches that failed to meet established specification(s) and their investigation'. ICH Q10 goes further: a process performance and product quality monitoring system that provides assurance of continued capability, identifies areas for continual improvement and feeds the management review. FDA's 2011 guidance on process validation turned trending into a daily matter. Stage 3, continued process verification, asks the firm to collect and analyze product and process data during routine production, with statistical methods appropriate to the data, to detect undesired process variability and to set the frequency of sampling and monitoring from an understanding of the process. The data are meant to be evaluated by trained personnel against the process's demonstrated variability, not against the specification alone. What none of these say is which chart, which limits or which rules. A specification limit tells you whether the lot can be released; it tells you nothing about whether the process that made it is the same process that was validated. The chart and the rules are the firm's decision, written into a procedure, and the record of every signal and its disposition is the evidence that the decision was applied. What the regulations and guidance ask, and what the signal record shows Expectation Where it comes from What the signal detection record shows Periodic review of batch and laboratory results for trends 21 CFR 211.180(e); EU GMP Chapter 1 (product quality review) The parameter's chart with its limits, every signal since the last review and QA's disposition of each Ongoing monitoring of process performance against its demonstrated variability FDA process validation guidance, stage 3; ICH Q10 section 3.2.1 Alert and action limits from the process's own history, run rules on every new result, capability against specification Statistical methods appropriate to the data, applied by trained people FDA process validation guidance, stage 3 The method named on the record: I-MR chart, sigma from the moving range or the standard deviation, Nelson or Western Electric rules A decision on each trend and the action taken ICH Q10; ICH Q9(R1) risk-based decisions Investigate, monitor or not a signal, with the reason, the signature and the investigation linked Data integrity of the trend record FDA data integrity guidance Results as recorded or received through an interface, a hash-chained audit trail, signatures with password re-entry A reading of 21 CFR 211.180(e), EU GMP Chapter 1, ICH Q10 and FDA's process validation guidance, stage 3. The wording of each expectation is ours; the sources are listed at the end. ## The chart: centre line, sigma, alert and action limits Most GMP parameters arrive one value per lot or one value per day: an assay, a fill weight, a bioburden count, a tablet hardness, a chromatographic resolution. There is no rational subgroup to average, so the individuals and moving range chart is the right tool, and it is the one ASTM E2587 and the usual pharmaceutical trending procedures describe. The centre line is the mean of the points the limits rest on: all points, the first N as a baseline, or a historical mean and sigma the firm approved. Sigma is estimated from the average moving range divided by d2 (1.128), the within-process estimate that is not inflated by a shift already present in the data; the sample standard deviation is available when the procedure asks for it. Two sets of limits sit on the chart. The alert limits at two sigma catch the single result that is unusual but not alarming: in a stable process one result in twenty lands there by chance, so one alert is a note, two in three is a rule. The action limits at three sigma catch the result that a stable process produces about three times in a thousand: that is a special cause until shown otherwise. The specification limits sit on the same chart, usually well outside the action limits, so the reader sees the distance between what the process does and what the product needs. The moving range chart beside it, with its upper limit at D4 times the average moving range, catches the jump between two consecutive results that the individuals chart would miss. Capability turns that distance into a number. Cp and Cpk use the within-process sigma from the moving range and say what the process could do while in control; Pp and Ppk use the overall standard deviation and say what it has done. A Cpk of 1.33 means the nearer specification limit is four sigma from the centre; a Ppk well below Cpk means the process is capable but not stable, which is exactly the case run rules exist for. Twenty lots of an assay: in specification throughout, shifted from lot eleven Illustrative results, not a measurement. The specification is 95 to 105 percent of label claim. Every lot passes. The alert limit catches lot 13; the run rule catches the shift at lot 17, nine points above the centre line; without the chart the shift would have been found at the annual review. ## The run rules: what a shift, a drift and a mixture look like Single-point limits see the result that is far out. The patterns that precede an out-of-specification result are made of results that are each unremarkable, and the run rules are the agreed definitions of those patterns. Lloyd Nelson's eight tests of 1984 are the default in QLogix; the four Western Electric rules, with a run of eight rather than nine, are the alternative for a procedure that cites them. Rule 2, nine consecutive points on one side of the centre line, is a shift of the mean: a new raw material lot, a changed setting, a different operator or instrument. Rule 3, six points steadily rising or falling, is a drift: wear, fouling, a reference standard degrading. Rule 4, fourteen points alternating up and down, is over-adjustment or two alternating sources. Rules 5 and 6, two of three beyond two sigma and four of five beyond one sigma on the same side, are smaller shifts seen early. Rule 7, fifteen points inside one sigma, is stratification: either the limits are too wide or mixed data is hiding the variation. Rule 8, eight points beyond one sigma on either side, is a mixture: two populations on one chart, two tanks, two presses, two analysts. Each rule names the pattern, and the name is where the investigation begins. A shift sends the investigator to what changed at the first point of the run; a drift to what wears; a mixture to what is being charted together that should be charted apart. The chart also says what is not a signal: a result that passes every rule is noise, and recording that conclusion with a reason is part of the trending record. Nelson's eight rules, the pattern each names and the question each raises Rule Pattern What it usually means First question for the investigation 1 One point beyond three sigma A special cause on that lot What was different about this lot, instrument or analyst? 2 Nine points on one side of the centre line A shift of the mean What changed at the first point of the run? 3 Six points steadily rising or falling A drift What wears, fouls or degrades at this rate? 4 Fourteen points alternating up and down Over-adjustment, or two alternating sources Is the process being corrected after every result? Are two sources alternating? 5 Two of three beyond two sigma, same side A small shift, seen early Same as rule 2, with less data 6 Four of five beyond one sigma, same side A small shift, seen early Same as rule 2, with less data 7 Fifteen points within one sigma Limits too wide, or mixed data hiding variation Are the limits from the right period? Is this one process or several? 8 Eight points beyond one sigma, either side A mixture of two populations What is being charted together that should be apart? The default rule set on a QLogix signal detection record; the Western Electric set replaces rules 1, 2, 5 and 6 with their four tests. Every rule is evaluated on every new result. ## From a result to a dispositioned signal: the record A chart that lives in a spreadsheet is reviewed when someone opens it. A signal detection record is reviewed when a result arrives. The record holds one parameter: its source (the LIMS, a batch record, a monitoring system, a pasted or uploaded data set), its specification, its baseline and sigma method, its rule set and every result with its date and lot. Results come in as they are recorded, by hand, by file, or through the LIMS interface and the REST API, and each new result is judged against the limits and the rules at once. When a rule fires, the AI reads the signal in the context of the parameter: the kind (limit, outlier, shift, trend, variation), the severity and the confidence, what in the data supports it, what changed at the first point of the run as far as the linked records say, and what to check first. It writes an interpretation and a recommended disposition. It does not decide. QA records the disposition, investigate, monitor or not a signal, with a reason, and signs; an investigation opened from the signal is linked both ways, so the chart shows the investigation and the investigation shows the chart. The process state on the record (in control, shifted, trending, unstable, out of specification, insufficient data) follows from the open signals. The review cycle closes the record for a period: the reviewer and the approver sign that the parameter was trended, every signal was dispositioned and the process state is as stated. That signed record, one per parameter per period, is the trending evidence for the product quality review, and the dashboard shows which parameters have open signals now. What happens to 1,000 results on a trended parameter - Results recordedby hand, by file or through the LIMS interface1000 - Judged against limits and rules at oncealert, action, specification, Nelson's eight, moving range1000 - Rule firedmostly single alerts and short runs38 - Interpreted by the AIkind, severity, confidence, what to check38 - Dispositioned by QAnone left open at the period review38 - Monitorwatched for the next results21 - Investigatean investigation opened and linked9 - Not a signalwith the reason on record8 An illustrative model of one year on a mid-volume product, not a measurement. Each stage is a count of results. Almost everything is noise; the point of the record is that the few signals are each read, dispositioned and signed, and that the dispositions are kept with the chart. ## Setting it up without fooling yourself The limits are only as good as the period they come from. A baseline that includes a shift gives limits wide enough to hide the next one; a baseline of ten points gives limits that move with every result. Twenty to thirty points from a period the firm believes was in control, or a historical mean and sigma approved in the procedure, are the usual starting points, and the record says which was used and when the limits were last recalculated. Mixed data is the second trap. Two presses, two filling lines, two analytical columns or two suppliers charted together produce a wide, calm-looking chart that fires rule 8 or rule 7 and hides a shift on one of them. One parameter per stream is the rule, and a stream that must be charted together should be charted together on purpose, with that choice recorded. The third is the response. A rule that fires and is dispositioned 'monitor' forty times in a row is not being monitored. The record makes that visible: the dispositions are on the chart, the open signals are counted on the dashboard, and the alert rules say when a signal has waited too long for a decision. Which chart and which rules, by the kind of data Choice One result per lot, specification exists Several results per lot or per day Counts (particles, colonies, defects) A measured drift over time (stability) Chart I-MR chart X-bar and R, or I-MR on the lot mean I-MR on counts, or c or u charts Regression against time, not a control chart Sigma Moving range / d2 Range or standard deviation within subgroup Moving range; square root for Poisson counts Residual standard error Rules Nelson's eight Nelson's eight or Western Electric Rules 1 to 3; others fire often on counts Slope and intercept confidence, not run rules Capability Cp, Cpk, Pp, Ppk against specification Cp, Cpk, Pp, Ppk Rarely meaningful Shelf-life projection What the data is A guide, not a rule. The chart type and the rule set are the firm's decision in its trending procedure; the record names which was used. ## What changes for QA The annual product review stops being the first time anyone looks. Each parameter has a chart that has been watched all year, every signal on it has a disposition and a signature, and the review is a reading of decisions already made rather than a search for trends in a spreadsheet assembled in the last week of the quarter. Out-of-specification results fall, because the shift that produces them is seen with results still in specification, and the deviation is opened on a trend rather than on a failed lot. The investigations that are opened are better framed: the rule names the pattern, the AI names what changed at the first point of the run, and the investigator starts from a question instead of from a result. The figure is an illustrative model of where the hours go, not a measurement. The point is where they move: out of assembling charts and arguing about whether something is a trend, into the dispositions and the investigations that are the actual work. Where the trending hours go across a product's year - Assembling charts: spreadsheetsexports, formulas, screenshots120 h - Assembling charts: recordsresults arrive, charts update8 h - Deciding what is a signal: spreadsheetsargued at the review60 h - Deciding what is a signal: recordsdispositioned as they fire30 h - Investigating real signals: boththe work that should take the time90 h - Writing the product quality review: spreadsheetscharts rebuilt, trends re-argued40 h - Writing the product quality review: recordssigned dispositions read back12 h An illustrative model of one product with twelve trended parameters, not a measurement. Hours per year by activity, spreadsheet trending first, signal detection records second. ## Common questions ### Is a result beyond the alert limit a deviation? No. A single result between two and three sigma happens about one time in twenty in a stable process. It is a signal for QA to disposition: monitor, investigate or not a signal, with the reason. A result beyond the action limit at three sigma, or outside specification, is a special cause until shown otherwise and usually opens an investigation. ### Which run rules should a pharmaceutical trending procedure use? Nelson's eight tests or the four Western Electric rules are both accepted; what matters is that the procedure names the set, the record shows it was applied to every result, and the dispositions are kept. Nelson's rules are the default in QLogix because they name more patterns (drift, alternation, stratification, mixture) and each name points the investigation somewhere. ### How many points are needed before the limits mean anything? Twenty to thirty points from a period believed to be in control is the usual baseline for an individuals chart; fewer gives limits that move with every result. A historical mean and sigma approved in the procedure are an alternative when the history is long and stable. The record says which was used and when the limits were last recalculated. ### Does the AI decide whether a signal is real? No. The AI reads each signal that the rules raise, states the kind, severity and confidence, what in the data supports it and what to check first, and recommends a disposition. QA records the disposition with a reason and signs; an investigation opened from a signal is linked to it. The AI's reading stays on the record with the tool and time it was made. ### How does this relate to continued process verification? Continued process verification, stage 3 of FDA's process validation guidance, asks for ongoing collection and statistical evaluation of product and process data to detect undesired variability. A signal detection record per critical parameter, with results judged as they arrive and every signal dispositioned and signed, is that evaluation as a continuous record rather than a periodic report. ## Sources - [FDA, Process Validation: General Principles and Practices (January 2011)](https://www.fda.gov/media/71021/download) - [21 CFR 211.180 General requirements (records and reports), eCFR](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-211/subpart-J/section-211.180) - [EudraLex Volume 4, Chapter 1: Pharmaceutical Quality System](https://health.ec.europa.eu/system/files/2016-11/vol4-chap1_2013-01_en_0.pdf) - [ICH Q10: Pharmaceutical Quality System](https://database.ich.org/sites/default/files/Q10%20Guideline.pdf) - [ICH Q9(R1): Quality Risk Management](https://database.ich.org/sites/default/files/ICH_Q9%28R1%29_Guideline_Step4_2023_0126_0.pdf) - [ASTM E2587, Standard Practice for Use of Control Charts in Statistical Process Control](https://www.astm.org/e2587-16.html) - [Nelson, L.S., The Shewhart Control Chart: Tests for Special Causes, Journal of Quality Technology 16(4), 1984](https://doi.org/10.1080/00224065.1984.11978921) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Signal detection - SPC - Continued process verification - Trending ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality data](https://qlogix.ai/services/connected-quality-data) [All articles](https://qlogix.ai/blog) ## More from the blog - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [AI in quality · 14 October 2026Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Backlog triage by risk for QA teams | QLogix Source: https://qlogix.ai/blog/backlog-triage-risk-based-prioritization AI in quality # Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by age By QLogix · Published 15 October 2026 · 9 min read The short answer Most quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible. Typical cycle times in days for deviations, CAPAs and change controls: the yardstick age is measured against, and never a reason for priority on its own. 30 / 90 / 90 Factors weighed in order: patient and product impact, regulatory exposure, recurrence, due dates and commitments, age against cycle time, the firm's own classification. 6 Signature on the queue: QA approves the priorities, escalates the critical items, or sends the data back for correction. 1 ## Key takeaways - 21 CFR 211.192 requires that unexplained discrepancies and failures be thoroughly investigated and that the investigation extend to other batches that may be associated; 21 CFR 211.22 makes the quality unit responsible for those decisions. ICH Q9(R1) asks that the effort, formality and documentation of quality risk management be commensurate with the risk. A backlog worked oldest-first satisfies neither: it spends the same effort on everything and decides nothing. - Risk-based triage weighs, in order, patient and product impact, regulatory exposure, recurrence across items, due dates and commitments, age against the typical cycle time, and the classification the firm already gave. Age alone makes an item Medium at most; age with a risk factor raises it. - Recurrence is the factor a person working one item at a time cannot see: the same cause, equipment, product, area, supplier or operator in three open items raises all three and is itself a finding. Reading the whole backlog at once is what a triage pass adds. - The AI ranks and explains; QA decides. Every item gets a suggested priority, a score and a one-line rationale with its evidence; QA sets the actual priority with a reason and signs the queue. The signed record is the answer to the inspector's question about why the sterile deviation was not worked first, or why it was. ## Why oldest-first is the wrong order A quality backlog is a list of open deviations, CAPAs, change controls, complaints, OOS results and audit findings, each with an owner, a due date and a status. In most systems it is sorted by age, because age is the one field every record has and the one metric management asks about. So the team works the oldest items, the overdue count falls, and the deviation on a Grade A filling line that was opened last week waits behind a documentation query from the spring. The regulations do not ask for the oldest item first. 21 CFR 211.192 asks that failures and discrepancies be thoroughly investigated, with the investigation extended to other batches that may have been associated with the specific failure, and 21 CFR 211.22 makes the quality unit responsible for approving or rejecting what affects the product. ICH Q9(R1), revised in 2023, asks that the level of effort, formality and documentation of quality risk management be commensurate with the level of risk, and names subjectivity in risk assessments as the thing to manage. EU GMP Chapter 1 asks for a quality system that ensures deviations are investigated and that management reviews the system's performance. None of that is satisfied by a sorted list. The regulators' interest is in whether the firm knows which open items could affect a patient, a batch on the market or a regulatory commitment, and whether those are being worked first. That is a ranking question, and it has to be answered across the whole backlog at once, not one record at a time. What the regulations ask, and what a signed triage record shows Expectation Where it comes from What the triage record shows Failures and discrepancies thoroughly investigated, extended to associated batches 21 CFR 211.192 Items with product on hold, a batch disposition pending or an associated batch ranked above the rest, with the evidence quoted The quality unit decides what affects the product 21 CFR 211.22 QA's priority on every item, with a reason, and the queue approved by signature Effort and formality commensurate with the risk ICH Q9(R1) A ranking by impact, exposure, recurrence and dates; low items confirmed as low rather than ignored Deviations investigated and the system reviewed by management EU GMP Chapter 1 Critical items escalated to management review; the patterns across items named A record of the decision and its basis 21 CFR 211.180, 211.192(b) The ranking, the rationale per item, QA's changes with reasons and the signatures, hash-chained A reading of 21 CFR 211.22 and 211.192, EU GMP Chapter 1 and ICH Q9(R1). The wording of each expectation is ours; the sources are listed at the end. ## The factors, in the order they are weighed Triage weighs six things, in roughly this order. First, patient and product impact: sterile or aseptic processing in Grade A or B areas, product on hold or already distributed, a batch disposition pending, a potential for harm, complaints or adverse events linked. Second, regulatory exposure: reportable events, regulatory commitments and inspection responses with dates, overdue CAPAs from inspection findings, data integrity. Third, recurrence: the same cause, equipment, product, area, supplier or operator in more than one open item, which raises every item in the cluster. Fourth, due dates and commitments: overdue or due within two weeks, effectiveness checks due, change controls blocking a release. Only fifth comes age, measured against the typical cycle time for the kind of record: thirty days for a deviation, ninety for a CAPA, ninety for a change control. Age alone makes an item Medium at most. Age with a risk factor raises it, because an overdue deviation on a distributed batch is a different thing from an overdue deviation on a training record. Sixth is the classification the firm already gave, critical, major or minor, when it is present; it is evidence of the firm's own view and it is weighed, not copied. Where the impact is not stated, the item is ranked conservatively and flagged so QA checks it. An item whose record says nothing about the product, the batch or the area is a data problem before it is a risk problem, and the ranking says so rather than guessing. How the factors combine into a priority Strongest factor on the item Within the cycle time Past the cycle time Past twice the cycle time Patient or product impact (Grade A/B, batch pending, distributed) Critical Critical Critical, escalate Regulatory exposure (commitment, inspection finding, reportable) High Critical Critical, escalate Recurrence (same cause, equipment, area in several items) High for the cluster High Critical Due dates (overdue, blocking a release, effectiveness check due) Medium High High Age only, impact not stated Low Medium, impact to be checked Medium, impact to be checked Age against the cycle time A guide to the ranking, not a formula. The rows are the strongest factor on an item; the columns are its age against the typical cycle time for its kind. The AI suggests; QA sets the priority. ## From the export to the signed queue A triage record starts from the backlog as it is: an export from TrackWise, Veeva Vault, MasterControl or SAP, a register, a screenshot, a PDF report, or the firm's own open QLogix records pulled directly. The AI first turns the sources into a clean list, one entry per record with its reference, type, title, status, owner, dates, classification and whatever the fields say about product, batch and area, and reports what could not be read. The list is edited before anything is ranked: a wrong date is corrected, a closed item removed, a missing field filled. The ranking then reads the whole list at once. Every item gets exactly one entry with a suggested priority, a score and a one-line rationale; the risk factors that apply are named as short flags, each with the evidence quoted from the item's own fields or a document, and an empty quote is better than an invented one. The overview names the patterns across items: a recurring cause, a cluster of overdue CAPAs, records stalled in review, and what information was missing. Closed or cancelled items that slipped into the export get Low with the action 'confirm closed and remove'. QA works the ranking, not the list. Each suggested priority is kept or changed with a reason, the rationale edited where QA knows better, and the decision recorded: priorities approved and the queue worked in this order, the critical items escalated to management review, or a re-triage ordered once the backlog data is corrected. The review and approval cycle signs the queue, and the record keeps the export's fingerprint, the AI's ranking with its tool and time, every change QA made, and the signatures. A triage pass, step by step - 1 Bring the backlog in An export, a register, a report or the open QLogix records; documents attached with their fingerprints. - 2 Extract the items One entry per record: reference, type, title, status, owner, dates, classification, product, batch, area; what could not be read, listed. - 3 Correct the list Dates fixed, duplicates and closed items removed, missing fields filled by QA before ranking. - 4 Rank by risk Every item with a suggested priority, a score, a one-line rationale and its flags with quoted evidence; patterns across items named. - 5 Decide QA keeps or changes each priority with a reason, edits rationales, records the decision on the queue. - 6 Sign and work the queue Reviewer and approver sign; the queue is worked in priority order; the next triage pass starts from what is still open. Each step produces an audited change on one record. The AI's steps draft; QA's steps decide and sign. ## What a backlog of 120 items looks like after triage The figure is an illustrative model of a mid-sized site's open items, not a measurement. The shape is typical: most items are Low or Medium and stay in the queue to be worked in order, a minority carry a risk factor that puts them High, and a handful are Critical because a batch, a patient or a regulatory commitment is on the line. The Critical items are rarely the oldest ones. In the model, two of the five are under a month old, and three of the twelve items older than twice their cycle time are Low. The patterns are the second product of the pass. In the model, four deviations on two filling lines share a cause, a stopper supplier's lot, which none of the four owners could see from their own record; triage raises all four and names the cluster, and the cluster is itself a finding for the CAPA system. Three CAPAs from the same inspection are overdue together, which is a regulatory exposure rather than three late items. Each number in the figure is also a question the record can answer. Why is item 47 Critical? Because its flag says product distributed and the quote is the field that says so. Why was item 12, the oldest, left Low? Because QA wrote the reason when it kept the suggested priority, and signed. 120 open items through one triage pass - Items in the exportdeviations, CAPAs, change controls, complaints, OOS120 - Extracted and correctedsix closed or duplicate items removed114 - Ranked with a rationaleevery item, one entry, flags with evidence114 - Lowworked in order; nine older than twice the cycle time61 - Mediumage with a weak factor, or impact not stated34 - Highdue dates, recurrence cluster, regulatory commitment14 - Criticalbatch pending, product distributed, Grade A deviation5 - Priorities changed by QAeach with a reason, then signed11 An illustrative model, not a measurement. Each stage is a count of items. The point is the shape: a few Critical items that were not the oldest, a cluster found by reading across items, and nothing left without a decision. ## The queue after the signature: alerts, the next pass and the metrics A signed queue is worked, and it ages. Alert rules watch the records behind it: a critical investigation open past its lead time, a CAPA whose effectiveness check is due, an adverse-event complaint undecided, a record stalled in review. The next triage pass starts from what is still open, with the previous priorities and reasons on the earlier record, so the drift of an item from Low to High over three passes is visible and is itself a signal about the process. The metrics change meaning. An overdue count says how many items are late; a triage record says how many late items carry a risk factor and how many are late because they are low and were correctly left. Management review gets the second number, with the Critical items escalated by name and the patterns across the backlog as findings. That is the metric ICH Q10's management review asks for: not the size of the backlog but what in it matters and whether it is being worked. The cycle is deterministic where it should be and judged where it must be. The rules raise the alerts; the AI ranks and explains; QA decides and signs; the next pass reads the decisions. Nothing in it changes a record's priority without a person and a reason. The triage loop - 1 . CollectThe open items from the systems that hold them, or the open QLogix records directly. - 2 . RankImpact, exposure, recurrence, dates, age, classification; a rationale and flags with evidence per item; patterns across items. - 3 . DecideQA keeps or changes each priority with a reason and records the decision on the queue. - 4 . SignReviewer and approver sign the queue; the record keeps the export, the ranking and every change. - 5 . Work and watchThe queue is worked in order; alert rules watch the records behind it. - 6 . ReportCritical items and patterns go to management review; the next pass starts from what is still open. - A triage pass takes minutes on the records; the hours go into the decisions on the Critical and High items. Each turn reads the queue the previous turn signed. QA decides at every turn; the AI only ranks and explains. ## What changes for the QA lead The Monday question changes from 'what is overdue' to 'what matters'. The overdue list still exists, but the team works the sterile deviation first and can say why the labelling query waited. The recurrence that nobody could see from inside one record is found by reading across them, and the CAPA system gets a finding it would otherwise have met at the next inspection. The inspection conversation changes too. Asked why an item was open for a hundred days, the QA lead opens the triage record: the item's suggested priority, QA's priority and reason, the signatures and the date, and the later pass where it rose or stayed. The answer is on record, in the words of the person who decided. The figure is an illustrative model of where a QA lead's hours go in a month, not a measurement. The time out of sorting and arguing goes into the decisions on the items that carry risk, which is the work the regulations ask the quality unit to do. Where the QA lead's month goes - Building the list: spreadsheetexports merged, dates corrected12 h - Building the list: triage recordextracted and corrected on the record2 h - Deciding what to work: spreadsheetargued weekly, not recorded16 h - Deciding what to work: triage recordpriorities kept or changed with reasons6 h - Working the Critical and High items: boththe work that should take the time36 h - Explaining the order to management and inspectors: spreadsheetreconstructed from memory8 h - Explaining the order: triage recordthe signed record, read back1 h An illustrative model of one site, not a measurement. Hours per month by activity, a backlog worked oldest-first from a spreadsheet first, a signed triage queue second. ## Common questions ### Does the AI set the priorities? No. It suggests a priority, a score and a one-line rationale for every item, with the risk factors named and the evidence quoted from the item's own fields. QA keeps or changes each priority with a reason, records the decision on the queue and signs it. The AI's ranking stays on the record with the tool and the time it was made. ### Which systems can the backlog come from? Any that can export: TrackWise, Veeva Vault, MasterControl, SAP, spreadsheets, registers, screenshots and PDF reports are read as attached documents or pasted text, and the firm's open QLogix deviations, CAPAs and change controls are pulled directly. The extraction reports what it could not read, and the list is corrected before it is ranked. ### Why does age count for so little? Because age says how long an item has waited, not what it could do. An overdue deviation on a distributed batch and an overdue training record are both overdue. The ranking measures age against the typical cycle time for the kind of record (30 days for a deviation, 90 for a CAPA or a change control) and lets it make an item Medium at most on its own; a risk factor is what raises it. ### How often should a backlog be triaged? Each pass is a signed record, so the frequency is the firm's to set: weekly for a large or fast-moving backlog, monthly for a small one, and always before a management review or an inspection. The next pass starts from what is still open and shows how each item's priority moved across passes. ### Is a triage record a GMP record? Yes. It holds the backlog as received with its fingerprint, the extracted list and QA's corrections, the AI's ranking, every change QA made with its reason, the decision and the reviewer's and approver's signatures, in a hash-chained audit trail. It is the record of why the backlog was worked in the order it was. ## Sources - [21 CFR 211.192 Production record review, eCFR](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-211/subpart-J/section-211.192) - [21 CFR 211.22 Responsibilities of quality control unit, eCFR](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-211/subpart-B/section-211.22) - [ICH Q9(R1): Quality Risk Management](https://database.ich.org/sites/default/files/ICH_Q9%28R1%29_Guideline_Step4_2023_0126_0.pdf) - [ICH Q10: Pharmaceutical Quality System](https://database.ich.org/sites/default/files/Q10%20Guideline.pdf) - [EudraLex Volume 4, Chapter 1: Pharmaceutical Quality System](https://health.ec.europa.eu/system/files/2016-11/vol4-chap1_2013-01_en_0.pdf) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Backlog triage - CAPA - Deviations - Quality metrics ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) - [AI in quality · 14 October 2026Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Predictive maintenance for GMP equipment | QLogix Source: https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp AI in quality # Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replaced By QLogix · Published 15 October 2026 · 12 min read The short answer A preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control. Failures on one asset before a Weibull fit of the times between them says whether they are infant, random or wear-out. 4 The life by which 10 percent of units have failed: a preventive interval set there catches nine failures in ten for a High-criticality asset. B10 Signatures a GMP calibration needs: the second-person review of the readings, and QA's decision when a reading is out of tolerance. 2 ## Key takeaways - 21 CFR 211.67 and 211.68 require written maintenance and calibration programmes with records of what was done; 21 CFR 820.70(g) and EU GMP Chapter 3 ask the same of devices and of European sites. None of them says how often. The interval is the firm's decision, and the records the firm keeps are the evidence that it was the right one. - The maintenance history already holds the reliability figures: corrective and emergency work orders are failures, their downtime is unavailability, their labour is the corrective share. Mean time between failures, mean time to repair, availability and the failure trend fall out of the records without a sensor on the machine. - With four or more failures, a Weibull fit of the times between them gives the shape: below 1 the asset is failing young and a shorter interval will not help; near 1 failures are random and the plan fits; above 1.3 failures rise with age and the interval should shorten toward the B-life for the asset's criticality, or the replacement case should be written. - Calibration belongs in the same model. An out-of-tolerance reading is a failure of the measurement, with its own impact assessment on what was measured since the last good calibration, a second-person review of the readings and QA's signed decision. On a linked equipment file, it counts toward the asset's reliability and its lifecycle stage. ## What do the regulations require of maintenance and calibration? For drugs, 21 CFR 211.67 requires that equipment be cleaned, maintained and, where appropriate, sanitized at appropriate intervals, under written procedures, with records of maintenance, cleaning and inspection kept. Section 211.68 covers automatic, mechanical and electronic equipment: routinely calibrated, inspected or checked according to a written program designed to assure proper performance, with written records of those checks maintained. Section 211.160(b)(4) adds the calibration of instruments, apparatus, gauges and recording devices at suitable intervals, under a written program with specific directions, schedules, limits for accuracy and precision, and remedial action when limits are not met. For devices, 21 CFR 820.70(g) requires equipment to be maintained under schedules with the activities and dates documented, and calibration under 820.72 with the accuracy and precision limits and the remedial action stated. EU GMP Chapter 3 asks that equipment be designed, located and maintained to suit its intended purpose, that repair and maintenance not present a hazard to product quality, and that measuring, weighing, recording and control equipment be calibrated and checked at defined intervals by appropriate methods with records kept. Annex 15 ties the qualified state of equipment to its maintenance and calibration: a change that affects it goes through change control and, where needed, requalification. What the regulations do not say is how long the interval should be. 'Appropriate', 'suitable' and 'defined' leave the number to the firm, which makes the interval a decision that needs a rationale and the records that show the rationale held. That is where the maintenance history comes in: a firm that can show mean time between failures, availability and the trend for each asset has a rationale for its intervals, and a firm that cannot is relying on the manufacturer's leaflet. What the regulations ask, and what the records must show Expectation Where it comes from What the connected record shows Maintenance at appropriate intervals under written procedures, with records 21 CFR 211.67; 820.70(g); EU GMP Chapter 3 A preventive plan per asset with its tasks, intervals and rationale; every work order with its tasks, labour, parts and verification Calibration under a written program with limits for accuracy and precision and remedial action 21 CFR 211.68; 211.160(b)(4); 820.72 An approved calibration requirement per asset: interval and basis, points and tolerances, traceability, the out-of-tolerance rule Records of the checks, with the result and the person 21 CFR 211.68(a); 211.194 As-found and as-left readings judged by the system, the certificate, the reviewer's signature Remedial action when limits are not met 21 CFR 211.160(b)(4); 820.72(a) The impact assessment on what was measured since the last good calibration, QA's signed decision, the investigation linked The qualified state maintained through change control EU GMP Annex 15 A plan change with its reason in the audit trail; a change control where the qualified state is affected Data integrity of the records FDA data integrity guidance Attributable entries, a hash-chained audit trail, signatures with password re-entry A reading of 21 CFR 211.67, 211.68 and 211.160(b)(4), 21 CFR 820.70 and 820.72, and EU GMP Chapter 3 and Annex 15. The wording of each expectation is ours; the sources are listed at the end. ## One equipment file: CMMS, calibration and the quality system on the same record graph The reason predictive maintenance stays a slide at most sites is that the data lives in three places. The CMMS holds the work orders, the calibration system holds the certificates, and the quality system holds the deviations that the failures caused. Each knows one third of the asset's story, and the engineer who wants the whole of it exports three spreadsheets. By the time they are joined the quarter is over. A connected architecture treats the maintenance asset as a node with everything about it linked both ways: its preventive plan and work orders, the calibration records of the instruments and gauges mounted on it, the deviations opened when it failed, the change controls that altered it, the validation project that qualified it. A link is an audited object with a direction and a meaning, not a column in a spreadsheet, so the inspector follows it as easily as the engineer. The figure shows the flows in QLogix: the maintenance asset at the centre, work orders feeding it failures and downtime, calibration records feeding it out-of-tolerance events, both feeding the quality system when a result or a batch is affected, and the reliability model and the AI reading across all of it under the user's own permissions. The data integrity expectations apply across the graph as they do inside any one record. Each work order and each calibration is recorded by the person who did or received it, time-stamped by the system, judged by the system where there is a limit to judge against, reviewed by a second person where the regulation expects one, and kept with its history in a hash-chained audit trail. Nothing in the model below reads a figure that was not recorded that way. The equipment file: maintenance, calibration, quality and the reliability model on one record graph - data systems - AI - people decide and sign - connected systems The dots travel in the direction of the data. Work orders and calibration records feed the maintenance asset; out-of-tolerance results and failures open investigations in the quality system; change controls alter the plan; the reliability model and the AI read across the graph and write only drafts; every arrow is an audited link. ## Reading the maintenance history: MTBF, MTTR, availability and the trend The first figures need no statistics. Every corrective or emergency work order completed on an asset is a failure; the downtime recorded on it is unavailability; the labour booked to it is the corrective share of the asset's maintenance. From the first record to today, the mean time between failures is the operating time divided by the number of failures, the mean time to repair is the mean downtime per failure, and the availability over the last year is one minus the downtime over the hours in the window. The failure trend compares the last 180 days with the 180 before: rising, steady or falling. Operating time is the figure most sites get wrong. A filling line that runs one shift and a bioreactor that runs continuously should not be compared in calendar days. An asset's operating hours per day, or its run-hour counter read on a known date, convert the history into operating time, and the MTBF comes out in hours the equipment actually ran. Without it the model says so and reports calendar days, which overstate the life of equipment that sits idle. These four figures are enough to see the fleet. The asset with an availability under 90 percent and a corrective share over 60 percent is already a replacement case, whatever its age. The asset whose recent failures are double the previous period's is wearing out. The asset with two failures in its first year, both within ninety days of installation, is in early life, and a shorter preventive interval will not fix an installation fault. Failures per quarter on one asset, and the point at which the trend becomes visible An illustrative count of corrective and emergency work orders per quarter on one asset over five years, not a measurement. The band is the rate the plan was written for; the run of rising quarters is what the trend rule and the Weibull shape pick up two quarters before the quarter that breaks the limit. ## When is an asset wearing out? The Weibull shape and the bathtub curve Counting failures tells you how many; the times between them tell you why. The Weibull distribution describes the three regimes of the bathtub curve with one shape parameter, beta. Below 1, the failure rate is falling with age: the asset fails young, from installation, commissioning, parts quality or operator error, and gets more reliable as those are fixed. Near 1, failures are random, independent of age, and a preventive interval cannot reduce them; condition monitoring and spares can. Above 1, the failure rate rises with age: wear, fatigue, fouling, drift. That is the regime a preventive interval is for, and the further beta is above 1, the sharper the knee and the more a well-placed interval saves. The fit needs data: with fewer than four failures the times between them cannot be ranked meaningfully, so the model holds its advice to the MTBF and the trend until there are enough. With four or more, median-rank regression on the log of the times gives beta and the characteristic life eta, the time by which 63 percent of units have failed. From those, the B-life for any probability follows: B10 is the time by which 10 percent have failed. A preventive task placed at B10 lets through one failure in ten; at B5, one in twenty. The criticality sets which B-life the interval targets. In QLogix a Critical asset's preventive interval is set at B5, a High asset's at B10, a Medium asset's at B20 and a Low asset's at B30; without a Weibull fit, the same idea is a fraction of the MTBF. The lifecycle stage follows from beta and the trend, then from availability and the corrective share: early life, useful life, wear-out, or end of life when wear-out comes with lost availability or a maintenance load that is mostly corrective. From the work orders to the lifecycle stage - 1 Failures and downtime Corrective and emergency work orders completed on the asset, and out-of-tolerance calibrations on its linked instruments; the downtime and labour recorded on them. - 2 Operating time Hours per day or the run-hour counter convert calendar days to operating hours; without them the model says it is working in calendar days. - 3 MTBF, MTTR, availability, trend Operating time per failure, mean downtime per failure, one minus downtime over the last year, the last 180 days against the 180 before. - 4 Weibull fit With four or more failures: beta (the shape) and eta (the characteristic life) from the times between failures. - 5 Lifecycle stage Early life, useful life, wear-out or end of life from beta, the trend, availability and the corrective share. - 6 Prediction and interval Next failure from the last one plus the median life; the preventive interval at the B-life for the criticality; a flag when the next task falls after the predicted failure. Each step is computed from the records as they stand. The engineer confirms before a plan changes; a change to the interval of a qualified asset goes through change control. ## Calibration belongs in the same model A calibration that finds a reading out of tolerance is a failure of the measurement, and in GMP it carries a heavier consequence than a seized bearing: everything the instrument measured since its last good calibration was measured wrong. The calibration module treats it as the regulations do. The reading is judged against the approved tolerance as it is entered, in the point's own unit or a convertible one. An as-found reading out of tolerance can be recorded as a pass after adjustment or a fail, never as a pass. The event opens the impact assessment, which the AI drafts from the readings, the requirement, the earlier calibrations and the linked records: the period at risk, the worst error, what the asset measured, the effect, the recommended decision. A person other than the one who recorded the calibration takes the decision and signs it with their password, and the investigation is linked where a result, a batch or a validated state may be affected. Two signatures make the calibration record inspection-ready. The second-person review of each calibration checks the readings against the certificate, the standards' certificates and their expiry, and the result, and signs Reviewed, or returns it with what must be corrected. Until it is reviewed, the calibration does not count: the asset shows awaiting review, and its due date rests on the last reviewed one. QA's signature on the out-of-tolerance decision is the second. Both go to the record's signatures with the record's hash at the time, under 21 CFR Part 11, without locking a record that must stay open for the asset's life. On a linked equipment file, each of those events counts toward the maintenance asset's reliability. A pressure transmitter that drifts out of tolerance every second calibration is telling the engineer the same thing a pump that trips every second month is: the asset is wearing, and the interval, the part or the asset should change. What happens to 100 calibrations - Calibrations recordedreadings judged as entered100 - Reviewed and signed by a second personthe rest returned and re-recorded96 - In tolerance as founddue date set from the interval88 - Out of tolerance as foundimpact assessment opened12 - Impact assessed, QA decision signednone left without a decision12 - Investigation openeda result or batch may be affected4 - Interval or asset changedthrough change control3 An illustrative model of a mid-sized site, not a measurement. Each stage is a count of calibrations. The losses a connected record removes are at review (nothing counts until a second person signs) and at the out-of-tolerance decision (nothing closes until QA signs with the impact assessed). ## From prediction to the maintenance plan: alerts, advisory and change control A prediction that stays on a dashboard changes nothing. Two alert rules carry it to the people who act. When an asset's failure history predicts the next failure within the rule's lead time, the asset's owner is told, with the figures and the advice. When the asset's next preventive task is scheduled after the predicted failure, or no task is scheduled, the rule says so, with the recommended interval. Both are deterministic: they read the same model as the dashboard, every quarter hour, and raise one alert per asset until it is resolved. The AI advisory reads the fleet's figures and the failure codes and drafts what a reliability engineer would propose: the preventive plan changes with the number of days and the task, the replacement and overhaul cases with the business case, the spares to stock and the training to give from the failure codes and repair times, and the questions the records leave open, such as the operating hours nobody recorded. It is decision support and it is labelled so. Nothing changes until the engineer takes it to the asset's plan with a reason, and to change control where the qualified state is affected. That last step matters for the inspector. An interval shortened because the Weibull shape rose above 1.3 is a defensible decision only if the record shows the figures, the reason, who decided and when, and the requalification where it was needed. On the connected record graph it does, because the plan change is an audited edit on the asset, the change control links to it, and the next calibration and the next work order carry the new interval forward. The predictive maintenance loop - 1 . RecordWork orders with downtime and labour; calibrations with readings reviewed and signed; operating hours or counters. - 2 . ComputeMTBF, MTTR, availability, trend and the Weibull fit per asset; the lifecycle stage; the predicted failure; the recommended interval. - 3 . AlertAssets predicted to fail within the lead time, and preventive tasks that fall after the predicted failure. - 4 . AdviseThe AI drafts plan changes, replacement cases, spares and training, and the questions the records leave open. - 5 . DecideThe engineer changes the plan with a reason, raises the change control, budgets the replacement. - 6 . VerifyThe next work orders and calibrations show whether the failures stopped; the model reads them on the next turn. - The loop runs on the records the site already keeps. No sensor is required to start; sensors make the operating time exact. Each turn of the loop reads the records the previous turn produced. The engineer decides at the plan step; the loop never changes a plan on its own. ## What changes for the site The intervals stop being folklore. Each asset's interval has a rationale the records support, and the rationale is updated when the records change. The reactive share falls, because wear-out assets get their intervals shortened before they fail rather than after, and the useful-life assets stop being over-maintained on a schedule copied from a leaflet. The replacement cases arrive with figures, which is what finance asks for. The calibration programme stops being a filing exercise. Due dates are kept by the system from reviewed calibrations, out-of-tolerance results cannot be closed without an assessed and signed decision, and the inspector follows the chain from the certificate to the batch the instrument measured without a spreadsheet. The figure is an illustrative model of where the hours go before and after, not a measurement. The point is where the hours move: out of gathering and firefighting, into the work that should take the time, which is deciding what to do with the asset. Where the maintenance engineer's month goes - Gathering data: disconnectedthree exports, one spreadsheet24 h - Gathering data: connectedthe dashboard, as it stands2 h - Firefighting breakdowns: disconnectedreactive share above 60 %36 h - Firefighting breakdowns: connectedwear-out caught early18 h - Deciding intervals and replacements: boththe work that should take the time16 h - Writing the change control: disconnectedassembled from emails8 h - Writing the change control: connecteddrafted, cited, final-reviewed3 h An illustrative model of a typical site, not a measurement. Hours per month by activity, disconnected systems first, connected record graph with the reliability model second. ## Common questions ### Does predictive maintenance need sensors or an IoT platform? No. The model described here reads the work orders, their downtime and labour, the calibration results and the operating hours a site already records. Sensors and counters make the operating time exact and let the model work in operating hours rather than calendar days; they are an improvement, not a prerequisite. ### Can the model change a preventive interval on its own? No. It computes the figures, predicts the next failure, recommends an interval and raises alerts. The maintenance engineer changes the plan with a reason, and a change to the interval of a qualified asset goes through change control and, where needed, requalification under EU GMP Annex 15. The AI advisory is a draft labelled as such. ### How many failures does the Weibull fit need? Four or more on one asset, so the times between failures can be ranked. Until then the model reports the MTBF and the trend and recommends the interval as a fraction of the MTBF by criticality. With the fit, the interval targets the B-life for the criticality: B5 for Critical assets, B10 for High, B20 for Medium, B30 for Low. ### Why does a calibration need a second-person review and a QA signature? 21 CFR 211.68 and 211.160(b)(4) require records of calibration checks and remedial action when limits are not met, and FDA's data integrity guidance expects records to be reviewed. A second person checks the readings against the certificate and signs Reviewed before the calibration counts for the asset's service; when a reading is out of tolerance, QA signs the decision on the impact to what was measured. Both signatures carry the signer, the meaning and the record's hash at the time. ## Sources - [21 CFR 211.67, Equipment cleaning and maintenance (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.67) - [21 CFR 211.68, Automatic, mechanical, and electronic equipment (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.68) - [21 CFR 211.160, General requirements for laboratory controls (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.160) - [21 CFR 820.70, Production and process controls (eCFR, current)](https://www.ecfr.gov/current/title-21/section-820.70) - [21 CFR Part 820, Quality System Regulation (eCFR, current)](https://www.ecfr.gov/current/title-21/part-820) - [EudraLex Volume 4, EU Guidelines for Good Manufacturing Practice (Chapter 3 and Annex 15)](https://health.ec.europa.eu/medicinal-products/eudralex/eudralex-volume-4_en) - [FDA, Data Integrity and Compliance With Drug CGMP: Questions and Answers, guidance for industry (2018)](https://www.fda.gov/media/119267/download) - [FDA, Part 11, Electronic Records; Electronic Signatures: Scope and Application](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/part-11-electronic-records-electronic-signatures-scope-and-application) - [FDA, Q9(R1) Quality Risk Management](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q9r1-quality-risk-management) - [ISPE, GAMP 5 Guide, Second Edition](https://ispe.org/publications/guidance-documents/gamp-5-guide-2nd-edition) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - CMMS - Calibration - Reliability - Equipment lifecycle ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [Connected quality data](https://qlogix.ai/services/connected-quality-data) - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 14 October 2026Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens complianceMost OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign.](https://qlogix.ai/blog/lims-investigation-capa-integration-ai) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # LIMS to investigation and CAPA, with AI | QLogix Source: https://qlogix.ai/blog/lims-investigation-capa-integration-ai AI in quality # Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens compliance By QLogix · Published 14 October 2026 · 12 min read The short answer Most OOS investigations lose their first days to gathering: the result, the raw data, the instrument's calibration, the standard's expiry, the analyst's history, the trend. When the LIMS, the investigation and the CAPA are one connected record graph, that evidence arrives with the result, AI reads it and drafts the hypotheses, and the effectiveness check is proven by the results that follow. The analyst, the investigator and QA still decide and sign. Phases in FDA's OOS investigation guidance: the laboratory phase first, then the full-scale investigation of manufacturing and sampling. 2 ALCOA+ attributes a laboratory record must keep, from attributable to available, across every system the result passes through. 9 Original result. An OOS result stands until an investigation finds an assignable cause; a passing repeat never replaces it. 1 ## Key takeaways - The laboratory is where most investigations start, and the regulations are specific: 21 CFR 211.160 and 211.165 require scientifically sound specifications and complete laboratory records, 211.192 a thorough investigation of any failure, and 211.194 the data derived from every test. FDA's OOS guidance sets the two-phase shape that inspectors expect. - A LIMS connected to the investigation and CAPA records is a different thing from a LIMS with an export button. The OOS result opens its investigation with the determinations, raw data fingerprints, instrument calibration state, standards and analyst attached and linked both ways; the investigation's findings govern what the laboratory may do next; the CAPA's effectiveness check reads the results that come after it. - AI accelerates the parts that are slow and repetitive: reading the trend, listing the testable hypotheses from the data around the result, drafting the phase one checklist and the write-up with every quote sourced. It proposes; the investigator tests and decides. - Compliance improves because the record is complete by construction: every result keeps its history, every link is audited, no result is tested into compliance, and the AI's drafts are reviewed and signed by people. The inspector sees one trail from the chromatogram to the closed CAPA. ## What do the regulations ask of laboratory results and their investigation? The laboratory controls in 21 CFR 211 are among the most specific in the regulation. Section 211.160 requires that specifications, standards, sampling plans and test procedures be scientifically sound and that laboratory controls include the calibration of instruments at suitable intervals. Section 211.165 requires testing of each batch against its specifications before release, with the accuracy and reliability of the test methods established. Section 211.194 lists what laboratory records must contain: the sample, the method, the weights and measures, the data derived from every test including all graphs and charts, the calculations, the results and how they compare with the acceptance criteria, and the initials of the person who performed each test and of the person who reviewed it. When a result fails, 211.192 takes over: a thorough investigation, extended to other batches that may have been associated with the failure, with a written record of the conclusions and follow-up. FDA's guidance on investigating out-of-specification results gives that investigation its shape for laboratory results: a phase one laboratory investigation that looks for an assignable laboratory error before any retesting, and a phase two full-scale investigation of manufacturing and sampling when the laboratory is not the cause, with retesting and resampling under a protocol and the original result never discarded without an assignable cause. EU GMP Chapter 6 asks the same of quality control, including that out-of-specification and out-of-trend results are investigated and that trend analysis is performed on results. FDA's data integrity guidance adds what a system must do with all of this: the data must be attributable, legible, contemporaneous, original and accurate, with audit trails reviewed, and nothing deleted or overwritten. MHRA's GxP data integrity guidance sets the same expectations across the lifecycle of the data. The thread through all of it is that the result, its raw data, the investigation it triggered and the action that followed must be one connected, reviewable story. What a laboratory result must carry with it Expectation Where it comes from What the connected record shows The specification the result is judged against, scientifically sound and approved 21 CFR 211.160; EU GMP Chapter 6 The approved specification version linked to the sample, with the method and limits Instruments calibrated at suitable intervals; standards and reagents in date 21 CFR 211.160(b)(4), 211.194(d) Calibration state and standard expiry checked at the time of analysis and kept with the result Complete laboratory records: data, calculations, comparison with criteria, who tested and who reviewed 21 CFR 211.194(a) Determinations as typed, raw data fingerprinted, calculated by the system, second-person review signed A thorough investigation of any failure, extended to other batches 21 CFR 211.192 The investigation opened from the result, pre-filled, linked both ways, with the batches assessed Phase one laboratory investigation before retesting; original result kept FDA OOS guidance (2022) No repeat without the investigation linked; invalidation only with an assignable cause and a reason Out-of-trend results investigated; trends reviewed EU GMP Chapter 6 Control chart with run rules on every test; out-of-trend flagged on the sample Attributable, contemporaneous, original, audit trails reviewed FDA and MHRA data integrity guidance Every entry under the analyst's own account, time-stamped, in a hash-chained audit trail A reading of 21 CFR 211.160, 211.165, 211.192 and 211.194, FDA's OOS guidance and EU GMP Chapter 6. The wording of each expectation is ours; the sources are listed at the end. ## What does a connected LIMS, investigation and CAPA architecture look like? A LIMS that stands alone produces results, and the quality system receives them by export, email or re-typing. Every hand-off loses context: the investigator gets the number but not the determinations, the instrument but not its calibration history, the analyst's name but not the other results they recorded that day. The connected architecture treats the laboratory record, the investigation and the CAPA as nodes in one record graph. A link is a first-class, audited object with a direction and a meaning: this investigation runs under this OOS result; this CAPA addresses this investigation's root cause; this change control implements this CAPA's action on the method. In QLogix the graph has a few more nodes. Specifications are controlled documents, so a result is judged against an approved version and a specification change goes through change control. Instruments and reference standards are records with their own calibration and expiry history, so a failed calibration lists the results recorded on that instrument since the last good one and opens a deviation investigation pre-filled with every sample to assess. Interfaces are controlled too: an instrument data system or an ERP connects through a registered interface that starts in validation, is bound to an API key with a laboratory scope, and is only heard once its validation reference is recorded. A result received through an interface is checked exactly like a typed one. The figure shows the flows. Data moves from instruments into the LIMS and from the LIMS into the investigation; findings move back as invalidations and repeat approvals; actions move into the CAPA and change control; every result and every link lands in the audit trail; and the AI reads across all of it, in the same permission model as the person asking. Nothing in the diagram is a file transfer. Each arrow is a link on a record that an inspector can follow in either direction. One record graph: laboratory data, investigation, CAPA and the AI that reads across them - data systems - AI - people decide and sign - connected systems The dots travel in the direction of the data. Instruments and the ERP connect through validated interfaces; the LIMS, investigation, CAPA and change control are records linked both ways; the AI reads the graph under the user's permissions and writes only drafts; people decide and sign; every arrow is an audited link. ## From an OOS result to an open investigation in one step The moment that matters is the one just after the result fails. In a disconnected laboratory the analyst tells a supervisor, someone opens a deviation form, and the investigator begins by asking for things: the chromatogram, the sequence, the standard's certificate, the instrument's last calibration, the analyst's training record, the results of the other samples in the run. Days pass before the first hypothesis is tested, and the phase one laboratory investigation that the guidance wants done promptly is done late and from memory. In the connected architecture the result's failure is a system event. The sample shows the result in red and offers to open the OOS investigation. The investigation opens pre-filled: the specification and limit, the determinations as typed, the fingerprints of the raw data files, the instrument with its calibration state at the time of analysis, the standards with their expiry, the analyst, and the other results in the same run. It is linked to the sample both ways. From here the rules are enforced by the system rather than by vigilance: a repeat can be recorded only under the linked investigation with the step it follows; the original result cannot be corrected into compliance; a passing repeat never replaces the failing one; and the sample goes to review only once the investigation that invalidated a result has itself been approved. The same pattern applies beyond the single result. A failed calibration lists every result recorded on that instrument since its last good calibration and opens a deviation investigation with all of those samples attached. A trend crossing a run rule on the control chart flags the sample as out of trend, which EU GMP Chapter 6 also expects to be investigated. Each path lands in the same investigation record with the same links. The OOS path, with the system enforcing the rules - 1 Result fails Judged by the system against the approved specification; the sample shows it in red and cannot be cancelled while the result stands. - 2 Investigation opens, pre-filled Result, determinations, raw data fingerprints, instrument calibration, standards, analyst and the rest of the run, linked both ways. - 3 Phase 1: laboratory Assignable laboratory error sought first. AI drafts the checklist from the data; the analyst and supervisor answer and sign. - 4 Phase 2: full scale Manufacturing and sampling when the laboratory is not the cause. Retests and resamples under a protocol, each repeat linked to its step. - 5 Disposition Invalidation only with an assignable cause and a reason; the original result stays on the record with its history. - 6 CAPA and effectiveness Actions address the confirmed cause; the effectiveness criterion is a result window and threshold the LIMS reports against. Each step is a record state, not a form. The investigation carries the laboratory data from the moment it opens; the laboratory cannot act on the result without the investigation; the CAPA's effectiveness is read from the results that follow. ## Where does AI accelerate the investigation? AI is not needed to judge a result against a limit; the system does that deterministically, and it must. AI earns its place in the parts of an investigation that are slow because they are reading and reasoning over scattered data. Given the pre-filled investigation and read access to the laboratory records, it drafts the phase one checklist specific to this result: the dilution and weighing checks, the sequence and integration review, the system suitability, the standard preparation. It lists hypotheses as testable statements, each with the data that would confirm or refute it and a quote from the record it came from: the standard was prepared two days before expiry; the instrument's last calibration was borderline on linearity; three other samples in the run drifted upward. It also reads the trend. The control chart on every test, with limits from the moving range and Nelson run rules, is a mechanical calculation; interpreting it is not. The AI's draft notes that the OOS point was preceded by six rising results that already triggered a run rule, that the shift coincides with a new lot of the reference standard, and that the other tests on the same instrument show no such shift, so the standard is a stronger hypothesis than the instrument. That is the kind of synthesis an experienced investigator does in an afternoon and a new one takes a week to reach. The limits are explicit. The AI proposes and cites; it does not set the root cause, invalidate a result, approve a repeat or close an investigation. Every draft is a draft on the record, with the model version named, until an investigator edits it, records the test results and signs. The AI's own final review before the record is routed for approval checks the investigation against what 211.192 and the OOS guidance expect, and lists what is still unsupported. Reading the trend around an OOS result An illustrative assay series for one test of one specification, drawn as the LIMS control chart draws it. The run of rising results before the failing point is the signal the AI points at: it was visible six samples earlier, and it coincides with the change of reference standard lot. The specification limits are the record's; the run rules and the hypothesis are the draft's. ## Data analysis at the speed of the laboratory The deeper gain is in what happens to the population of OOS results over a year, not to any single one. A laboratory that investigates each result from scratch finds assignable laboratory causes in some, confirms manufacturing causes in others, and leaves a residue of investigations that end in no assignable cause because the data to decide was never gathered or had been overwritten. A connected laboratory changes the shape of that funnel. Phase one is faster and more often conclusive because the checklist is specific and the data is present. Phase two extends to other batches because the system knows which batches were tested on that instrument, with that standard, by that analyst, in that period. The CAPA is tied to a cause rather than to a symptom because the investigation showed its evidence. The figure is an illustrative model, not a measurement, and the point is the shape rather than the numbers: losses in a disconnected laboratory are concentrated at the gathering and the no-cause stages, and those are exactly the stages a connected record graph and an AI reader remove. What remains is the scientific work: testing the hypothesis, deciding, signing. The same data serves continued process verification. Right-first-time rates, turnaround from receipt to approval, out-of-trend and invalidation counts by test, analyst and instrument are laboratory indicators the LIMS reports from the records themselves, and the AI can be asked, in plain language, what changed in the assay trend for a product over the last quarter and whether any open CAPA already addresses it. What happens to 100 OOS results - OOS resultseach opens its investigation100 - Phase 1 complete within targetdata present from the start88 - Laboratory cause confirmedresult invalidated with a reason34 - Phase 2 with batches extendedthe system lists them54 - Assignable cause foundevidence on the record78 - CAPA addressing the causeeffectiveness read from results46 - No assignable causemost probable cause stated22 An illustrative model of a mid-sized QC laboratory, not a measurement. Each stage is a count of results. In a disconnected laboratory the losses sit at gathering (investigations that stall) and at the no-assignable-cause outcome; the connected graph and the AI reader act on exactly those stages. ## Closing the loop: CAPA effectiveness proven by laboratory data A CAPA fails most often at its effectiveness check, which is written at closure and shows whatever happened to happen. The connected graph makes the check concrete at the time the plan is approved: the criterion is a test, a product, a window of results and a threshold. No OOS or OOT result for the assay on the next twenty batches; the capability index for the dissolution test above the target for three months; no invalidation attributed to standard preparation in the next quarter. Those are LIMS queries, and the CAPA record reads them. When the action changes the method, the standard preparation or the specification, it goes through change control, and the change control links back to the CAPA and forward to the new specification version, which the LIMS then judges results against. The same is true when the action is a system change: the interface to an instrument data system is suspended with a reason during the upgrade, revalidated, and made active again with the validation as the reason. Nothing in the loop depends on someone remembering to update a spreadsheet. Where time is saved, it is saved in the places the figure shows. The model is illustrative; the stages are the ones investigators describe when asked where the days go. The connected graph and the AI reader remove most of the gathering and the waiting and shorten the writing; they do not shorten the testing, which is the part that should take the time it takes. Where the days go in an OOS investigation - Gathering: disconnectedrequests to the lab, IT and QA6 days - Gathering: connectedattached when the investigation opens0.5 days - Phase 1 draft: disconnectedfrom memory and the SOP4 days - Phase 1 draft: connectedAI drafts from the data; investigator edits1 days - Testing hypotheses: boththe work that should take the time5 days - Write-up: disconnectedassembled from emails4 days - Write-up: connecteddrafted, cited, final-reviewed1.5 days An illustrative model of a typical investigation, not a measurement. Calendar days by activity, disconnected laboratory first, connected graph with AI second. Testing the hypotheses is deliberately unchanged: that is the scientific work. ## Compliance by design: what the inspector sees The compliance gain is not that AI writes better paragraphs. It is that the record is complete by construction. Every result is under the analyst's own account, time-stamped, calculated by the system, with the raw data fingerprinted and the calibration and expiry checks made at the time of analysis. Every change keeps the old value, the new value, the reason and the person in a hash-chained audit trail whose integrity is verified when the record is opened. Every link between the sample, the investigation, the CAPA and the change control is itself an audited entry with a direction and a reason. An inspector asking for the story of a failed batch gets one trail, from the chromatogram to the closed CAPA, in either direction. The AI sits inside that model, not beside it. It reads under the permissions of the person asking, in their organization only; it writes drafts that are labelled as drafts with the model version named; it never signs; and its use is itself audited. The assurance effort for the AI feature follows GAMP 5 Second Edition and FDA's computer software assurance guidance: a feature that drafts text for a person to review is lower risk than one that changes a record, and the effort is set accordingly, with a change of model going through change control. Part 11 and Annex 11 apply to the records and signatures the same way they did before the AI arrived. The layers below show how the responsibilities divide. The data layer is deterministic and validated: it judges, calculates, links and audits. The AI layer reads and drafts. The people layer decides and signs. Keeping those three apart is what lets a laboratory move faster without weakening a single control. Who does what in the connected laboratory - People decidePeople decide and signThe analyst records and corrects their own results; the second person reviews; the investigator tests hypotheses, sets the root cause and the CAPA; QA approves and signs. Every signature means something named. - AIAI reads and draftsPhase one checklist from the data; hypotheses with the evidence for each; the trend read; the write-up and the CAPA plan drafted with sources quoted; a final review against the regulations before routing. Drafts only, model version named, use audited. - DataData judges, links and auditsResults judged against the approved specification; calibration and expiry checked at analysis; raw data fingerprinted; OOS rules enforced; links between sample, investigation, CAPA and change control audited; hash chain verified on opening; interfaces validated and scoped. The data layer is deterministic and validated; the AI layer proposes and cites; the people layer decides and signs. The boundaries are enforced by the system, not by policy alone. ## Common questions ### Does connecting the LIMS to the investigation system change how OOS results must be handled? No. The connection enforces the handling FDA's OOS guidance already expects: a phase one laboratory investigation before any retesting, the original result kept, invalidation only with an assignable cause, retests and resamples under a protocol, and the investigation extended to other batches. What changes is that the system carries the data to the investigation and refuses the shortcuts, so compliance does not depend on vigilance. ### Can AI decide that a laboratory error caused an OOS result? No. The AI drafts the phase one checklist and lists hypotheses with the evidence for each; the analyst, the supervisor and the investigator test them, record the results and decide. An invalidation is recorded by a person with a reason, under the linked investigation, and the original result stays on the record. The AI's proposals are drafts with the model version named, and every edit to them is audited. ### What does an instrument or ERP interface need before the LIMS will accept its data? A registered interface that names the system, its version, its owner and what it exchanges; an API key scoped to the laboratory and bound to that interface; for an instrument data system, the instruments it may report for; and a recorded validation reference before the interface is made active. Every message is logged as accepted, rejected with a reason, or a duplicate, and a result received through an interface is checked exactly like a typed one. ### How is a CAPA's effectiveness proven from laboratory data? The effectiveness criterion is defined when the plan is approved as a test, a product, a window of results and a threshold: for example no OOS or OOT result for the assay across the next twenty batches. The CAPA record reads those results from the LIMS at the check, so the evidence is the laboratory's own data rather than a statement written at closure. ## Sources - [21 CFR 211.160, General requirements for laboratory controls (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.160) - [21 CFR 211.165, Testing and release for distribution (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.165) - [21 CFR 211.192, Production record review (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.192) - [21 CFR 211.194, Laboratory records (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.194) - [FDA, Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production, guidance for industry (revised 2022)](https://www.fda.gov/media/71001/download) - [FDA, Data Integrity and Compliance With Drug CGMP: Questions and Answers, guidance for industry (2018)](https://www.fda.gov/media/119267/download) - [EU GMP Chapter 6, Quality Control (EudraLex Volume 4)](https://health.ec.europa.eu/system/files/2016-11/2014-11_vol4_chapter_6_0.pdf) - [EU GMP Annex 11, Computerised Systems (EudraLex Volume 4)](https://health.ec.europa.eu/system/files/2016-11/annex11_01-2011_en_0.pdf) - [MHRA, GxP data integrity guidance and definitions (2018)](https://www.gov.uk/government/publications/guidance-on-gxp-data-integrity) - [FDA, Part 11, Electronic Records; Electronic Signatures: Scope and Application](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/part-11-electronic-records-electronic-signatures-scope-and-application) - [ISPE, GAMP 5 Guide, Second Edition](https://ispe.org/publications/guidance-documents/gamp-5-guide-2nd-edition) - [FDA, Computer Software Assurance for Production and Quality System Software (Federal Register notice, 2025)](https://www.federalregister.gov/documents/2025/09/24/2025-18468/computer-software-assurance-for-production-and-quality-system-software-guidance-for-industry-and) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - LIMS - OOS investigations - CAPA - Data integrity ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality data](https://qlogix.ai/services/connected-quality-data) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI for deviation investigations and root cause | QLogix Source: https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis AI in quality # AI in deviation investigations: faster, more consistent root cause analysis By QLogix · Published 14 October 2026 · 7 min read The short answer AI is most useful in an investigation at the start, where time is lost: it drafts the problem statement, the containment check, the 5 Whys and a fishbone whose causes are testable hypotheses, each quoting the write-up or an attached document. The investigator then tests, decides and signs. Nothing the AI writes becomes the root cause until a qualified person says so. Fishbone categories (the 6M) a complete analysis covers, each cause written as a testable hypothesis. 6 Phases in FDA's OOS investigation guidance: the laboratory phase, then the full-scale investigation. 2 Written record per investigation, with conclusions and follow-up, required by 21 CFR 211.192. 1 ## Key takeaways - 21 CFR 211.192 requires a thorough investigation of any unexplained discrepancy, extended to other batches that may be affected, with a written record of the conclusions and follow-up. EU GMP Chapter 1 asks for an appropriate level of root cause analysis and, where the true root cause cannot be determined, the most likely one. - AI is at its best drafting the parts investigators find slowest to start: the problem statement, the containment check, the 5 Whys and a fishbone whose causes are phrased as hypotheses with the test that would confirm or refute each, every quote naming its source. - An investigation that ruled out alternatives with objective evidence holds up at inspection; one that argued its way to a cause does not. The AI's job is to make the alternatives explicit and testable, not to pick one. - The investigator decides. The AI never sets the root cause, the classification or the CAPA; it proposes, cites and flags what is missing, and every edit to its draft is audited with a reason. ## What do the regulations expect from a deviation investigation? For drugs, 21 CFR 211.192 is the anchor: any unexplained discrepancy, or a batch that fails to meet its specifications, must be thoroughly investigated, whether or not the batch was distributed; the investigation extends to other batches of the same product and other products that may have been associated with the failure; and a written record must include the conclusions and the follow-up. EU GMP Chapter 1 adds that an appropriate level of root cause analysis is applied to deviations, suspected defects and other problems, and that where the true root cause cannot be determined, the most likely root cause is identified and addressed. FDA's guidance on investigating out-of-specification results, revised in 2022, sets the shape most firms use for laboratory results: a phase one laboratory investigation that checks for an assignable laboratory error before any retesting, then a full-scale phase two investigation of manufacturing and sampling when the laboratory is not the cause. ICH Q9(R1), adopted in 2023, asks that the formality and effort of risk-based work be commensurate with the risk, which applies as much to the depth of an investigation as to any other quality decision. In FDA's published inspection observation data, failures to thoroughly investigate discrepancies under 211.192, and inadequate CAPA procedures under the device regulation, have been among the most frequently cited observations year after year. The common thread in those citations is not a missing form; it is an investigation that did not go deep enough, did not extend far enough, or did not show its evidence. What an inspector looks for in an investigation record Expectation Where it comes from What the record must show Thorough investigation of every unexplained discrepancy 21 CFR 211.192 A problem statement, the facts gathered and the conclusion Scope extended to other batches and products that may be affected 21 CFR 211.192 Which batches were assessed and the result Appropriate level of root cause analysis; most likely cause when the true one cannot be found EU GMP Chapter 1, 1.4(xiv) The causes considered, the evidence for and against each, the conclusion Laboratory error ruled in or out before retesting FDA OOS guidance (2022 revision) Phase one checklist, analyst and supervisor review, retest justification Effort commensurate with risk ICH Q9(R1) The classification and why the depth of investigation fits it Written record with conclusions and follow-up 21 CFR 211.192 Signed conclusion, CAPA references, effectiveness follow-up A reading of 21 CFR 211.192, EU GMP Chapter 1 and FDA's OOS guidance. The wording of each expectation is ours; the sources are listed at the end. ## Where do investigations go wrong? - They start late. The first days after a deviation are when evidence is freshest, and they are often spent on the form rather than the facts. - The problem statement is vague. Without what, where, when and how much, every later step argues about the wrong thing. - They jump to human error. Retraining the operator closes the record and leaves the condition that let the error happen. - Causes are asserted, not tested. A fishbone with twenty causes and no tests is a brainstorm, not an analysis. - Scope stops at the batch in front of the investigator, when 211.192 expects other batches and products to be assessed. - The CAPA does not follow from the cause, so the effectiveness check, if there is one, cannot show anything. What happens to 100 deviations without hypothesis testing - Deviations openedevery record100 - Problem statement with what, where, when, extentspecific enough to test64 - Causes written as testable hypotheseswith a named test31 - Alternatives ruled out with objective evidencelab data, inspection, statistics22 - Root cause confirmed, not arguedholds up at inspection18 - Same cause recurs within a yearwhere it was not confirmed27 An illustrative model of a typical site, not a measurement: the figures show the shape of the funnel, where investigations lose their evidence and where recurrence comes from. Each stage is a count of deviations. ## What can AI draft, and from what? The AI reads the write-up and the documents attached to the record: the batch record, the SOP, the equipment log, the lab data, the previous investigation. From those it drafts the pieces of a structured investigation, and every statement it makes quotes the source it came from, by document and page, so a reviewer can check it in seconds. Where the documents do not support a statement, the draft says so and lists what is missing instead of filling the gap. The parts it drafts are the problem statement with its is and is-not boundaries, the containment and impact assessment, the 5 Whys chain, the Ishikawa fishbone across the six categories, a provisional root cause with a stated confidence, and a CAPA plan whose actions address the cause rather than the symptom. The investigator can change any line; each change is a separate audit entry with a reason, and a redraft replaces the earlier draft only with a reason too. The division of labour in an AI-assisted investigation - DataThe write-up and the attached documentsBatch record, SOP, equipment or maintenance log, lab record, training record, previous investigation, specification. - AIDraft investigation with sourced quotesProblem statement, containment and impact, 5 Whys, fishbone causes as hypotheses with tests, provisional root cause, CAPA plan, and the list of information still missing. - People decideTest, disposition, conclude, signInvestigators run the tests, record results, rule causes in or out, set the root cause and its confidence, approve the CAPA, and sign under 21 CFR Part 11. Data goes in, the AI drafts and cites, people test, decide and sign. The AI never changes a record on its own. ## How does hypothesis testing make the root cause robust? A potential cause becomes a hypothesis when it is written as a prediction: if this cause is responsible, the test will show this; if it is not, the test will show that. The AI drafts each fishbone cause that way, with a test method that prefers objective evidence (a laboratory analysis, an equipment inspection, an experiment, a statistical comparison of data) over interviews. The investigator records the result, and the record shows whether it supports, refutes or cannot decide the cause. When data exists, the comparison can be a statistical test: a t-test between two batches, a chi-square across shifts, a control chart that shows whether a shift happened when the suspected cause appeared. At the end, an AI review of the conclusion rates it as robust, adequate, weak or not established, says what an inspector would challenge, and lists the next steps to close the gap. The rating is advice; the conclusion and its confidence remain the investigator's. From potential cause to confirmed root cause - 1 Hypothesis If this cause is responsible, then this observable consequence follows. - 2 Test The method, preferring objective evidence: analysis, inspection, experiment, data comparison. - 3 Prediction What the test shows if the cause is responsible, and what it shows if it is not. - 4 Result The recorded outcome, with the data or the document it rests on. - 5 Disposition Supports, refutes or inconclusive; the root cause is the cause that survived and explains the facts. Every cause on the fishbone goes through the same five steps. A cause that cannot be tested is recorded as such, with the reason. ## How does the investigation stay compliant with AI in the loop? - Every quote names its source: the write-up, or an attached document by its ID and page. A reviewer can verify any statement without re-reading the whole file. - The AI's draft is a version of the record like any other. Replacing it needs a reason; every edit to it is an audit entry with who, when, what and why. - The audit trail records the AI tool and its version with each draft, so the record shows what produced the text that a person then reviewed. - Decisions stay with people: the dispositions, the root cause and its confidence, the classification, the CAPA and the signature. The AI cannot change a record by itself. - The tool is assured for its intended use: under GAMP 5 Second Edition and FDA's CSA guidance, a feature that drafts text for review sits at a lower process risk than one that changes a record, and a change of model goes through change control. Who does what, step by step Step AI drafts Person reviews Person decides Problem statement From the write-up Corrects boundaries Accepts the statement Containment and impact From records Checks scope Confirms actions Fishbone hypotheses and tests Six categories, with tests Adds or removes causes Chooses which to test first Running the tests Suggests the method Records results Performs and interprets Root cause and confidence Provisional, with evidence Reviews the verdict Sets the root cause CAPA plan Actions that address the cause Assigns owners and dates Approves the plan Classification and signature None None Classifies and signs Role Where the AI's contribution is high the person reviews; where the decision matters the person decides. The AI's share never reaches the signature. ## What changes in cycle time and in the quality of the conclusion? The first draft is where the time goes: gathering what the documents say, writing a problem statement that survives a second reading, and laying out the causes. With that draft on the screen within minutes of attaching the documents, the investigator's day starts at testing, not at a blank page. The same structure every time also makes investigations comparable, which is what a trend review and an inspector both need. The second gain is across records. A quality system that connects deviations, complaints, batch reviews, signals and CAPAs by shared lot numbers, equipment, products and cause wording shows the investigator the earlier records that share something with this one before the fishbone is drawn, so a recurrence is treated as one, and the new investigation starts from what the last one found. An investigation with AI support, day by day - Day 0Opened Deviation raised, documents attached Write-up, batch record, SOP and equipment log on the record. - Day 0AI draft Draft investigation Problem statement, containment, 5 Whys, fishbone with hypotheses and tests, sourced quotes, missing information listed. - Day 1In progress Hypotheses prioritised, tests assigned Investigator corrects the draft and assigns the laboratory and engineering tests. - Day 5In progress Results recorded, causes dispositioned Each cause marked supports, refutes or inconclusive, with its evidence. - Day 6Review Conclusion reviewed AI review rates the conclusion and lists what an inspector would challenge; investigator closes the gaps. - Day 8Signed CAPA approved, investigation signed Plan approval and electronic signatures; the effectiveness check is defined before closure. An illustrative sequence for a deviation with attached documents and laboratory follow-up. Days are working days from the deviation being raised. ## How QLogix supports investigations QLogix's Deviations and CAPAs product drafts the analysis from the write-up and attached documents with every quote sourced, keeps each fishbone cause as a hypothesis with its test and result, reviews the conclusion for robustness, connects the record to earlier ones that share a lot, equipment or cause, and keeps every change in a hash-chained audit trail with Part 11 signatures. The services pages below describe how we put it in place within an existing quality system. ## Common questions ### Can AI decide the root cause of a deviation? No, and it should not. The AI drafts the analysis and proposes a provisional root cause with its evidence; the investigator tests the hypotheses, records the results, sets the root cause and its confidence, and signs. Regulations place the investigation and its conclusion with the firm's qualified people, and an inspector will ask who concluded what and on what evidence. ### Does using AI in investigations need validation? The tool needs assurance for its intended use, as any software in the quality system does. Under GAMP 5 Second Edition and FDA's computer software assurance guidance the effort follows process risk: a feature that drafts text for a person to review is lower risk than one that changes a record. The assurance record names the model version, and a change of model goes through change control. ### What is the difference between a deviation, a non-conformance and an OOS result? A deviation is a departure from an approved instruction or standard during manufacturing or testing. A non-conformance is material, product or a process that does not meet its requirements. An out-of-specification result is a test result outside the acceptance criteria and is investigated first in the laboratory, under FDA's OOS guidance, before a full-scale investigation. ### How long should a deviation investigation take? Neither 21 CFR 211.192 nor EU GMP sets a number of days; they require a thorough, timely investigation whose depth fits the risk. Many firms set 30 calendar days as an internal target and document a justified extension when laboratory or engineering work takes longer. What inspectors challenge is an unexplained delay, not a documented one. ## Sources - [21 CFR 211.192, Production record review (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.192) - [FDA: Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production, guidance for industry (revised 2022)](https://www.fda.gov/media/71001/download) - [EudraLex Volume 4, Chapter 1: Pharmaceutical Quality System (2013)](https://health.ec.europa.eu/system/files/2016-11/vol4-chap1_2013-01_en_0.pdf) - [FDA: Q9(R1) Quality Risk Management, guidance for industry](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q9r1-quality-risk-management) - [FDA: Q10 Pharmaceutical Quality System, guidance for industry](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q10-pharmaceutical-quality-system) - [FDA: Inspection Observations (annual observation data)](https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/inspection-references/inspection-observations) - [FDA: Quality Systems Approach to Pharmaceutical CGMP Regulations, guidance for industry (2006)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/quality-systems-approach-pharmaceutical-current-good-manufacturing-practice-regulations) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Investigations - Root cause analysis - 21 CFR 211.192 - ICH Q9 ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI-assisted CAPA from root cause to effectiveness | QLogix Source: https://qlogix.ai/blog/ai-capa-management-effectiveness AI in quality # AI-assisted CAPA: from root cause to a verified effectiveness check By QLogix · Published 14 October 2026 · 7 min read The short answer A CAPA fails for structural reasons: actions that restate the symptom, no owner or date, an effectiveness check invented at closure, and no view across records to see the same cause coming back. AI helps most before the plan is approved: proposing actions that address the confirmed cause, drafting the effectiveness criterion up front, and matching the CAPA against earlier records. People approve, implement, verify and close. Elements of the ICH Q10 pharmaceutical quality system; the CAPA system is one of them. 4 QMSR in effect: ISO 13485's corrective and preventive action clauses now apply under 21 CFR Part 820. 2 Feb 2026 Things to keep apart in every CAPA: the correction, the corrective action and the preventive action. 3 ## Key takeaways - ICH Q10 names the CAPA system as one of four elements of the pharmaceutical quality system, fed by complaints, rejections, non-conformances, recalls, deviations, audits, inspections and trends, with effort commensurate with risk. - For devices, the QMSR, effective 2 February 2026, incorporates ISO 13485:2016, whose clauses on corrective and preventive action require actions to be taken without undue delay, their effectiveness verified, and the results recorded. - The commonest CAPA failures are structural: actions that treat the symptom, missing owners and dates, an effectiveness check written at closure to fit what happened, and no way to see the same cause across records. - AI's contribution comes before approval: actions that follow from the confirmed cause, an effectiveness criterion defined up front with its data and window, and a match against earlier records that share the lot, equipment or cause. Approval, implementation, verification and closure stay with people. ## What do ICH Q10, 21 CFR 211 and the QMSR require of CAPA? ICH Q10 describes the CAPA system as one of four elements of the pharmaceutical quality system, alongside process performance and product quality monitoring, change management and management review. It should be fed by complaints, product rejections, non-conformances, recalls, deviations, audits, regulatory inspections and findings, and trends, and it should use a structured approach to the investigation with the objective of determining the root cause, with the level of effort commensurate with the risk, as ICH Q9 describes. For drugs, 21 CFR 211.192 ties the CAPA to the investigation: the written record must include the follow-up. EU GMP Chapter 1 asks that appropriate corrective and preventive actions are identified and taken in response to investigations, and that their effectiveness is monitored and assessed. For devices, the Quality Management System Regulation took effect on 2 February 2026 and incorporates ISO 13485:2016 by reference, whose corrective and preventive action clauses require the firm to review non-conformities, determine causes, evaluate the need for action, plan and record the action, verify that it does not adversely affect the device, and review its effectiveness. CAPA requirements by framework Framework What it asks of a CAPA What the record shows ICH Q10 (drugs, PQS) A structured investigation to root cause; actions commensurate with risk; CAPA as an input to management review Source of the CAPA, root cause, actions, outcome, trend 21 CFR 211.192 (drugs) Investigation record with conclusions and follow-up Follow-up actions linked to the investigation EU GMP Chapter 1, 1.4(xiv) (drugs) Appropriate CAPA identified and taken; effectiveness monitored and assessed Actions, their completion, the effectiveness assessment QMSR with ISO 13485:2016 8.5.2 and 8.5.3 (devices) Review non-conformities, determine causes, plan, record and verify actions, review effectiveness Cause, action plan, verification, effectiveness review, records Our summary of the published texts; each document is in the sources. The QMSR's requirements are those of ISO 13485:2016 as incorporated. ## Why do CAPAs fail? - The action restates the symptom: retrain, re-clean, re-inspect. The condition that produced the event is untouched. - No owner or date, so the plan is approved but nothing is due to anyone. - The effectiveness check is written at closure, after the fact, and shows whatever happened to happen. - Correction, corrective action and preventive action are mixed together, so the record cannot show that the cause was addressed. - The CAPA is not connected to the deviations, complaints and signals that share its cause, so the same CAPA is opened again under a new number. - Closure is driven by the due date rather than by evidence that the action was implemented and worked. What happens to 100 CAPAs - CAPAs openedevery record100 - Actions address a confirmed causenot the symptom58 - Owner and due date on every actionsomeone is accountable74 - Effectiveness criterion defined at approvaldata, window, threshold36 - Implemented with evidence on filerecords, not statements61 - Verified effective at the checkcriterion met29 - Reopened or recurs under a new numberthe cost of the gaps21 An illustrative model of a typical site, not a measurement. The point is where the losses occur: at the plan, where the effectiveness criterion should be set, and at verification, where evidence is missing. Each stage is a count of CAPAs. ## Correction, corrective action, preventive action: why the distinction matters A correction deals with the non-conformity itself: the batch on hold, the rework, the replaced part. A corrective action removes the cause so it does not happen again. A preventive action removes the cause of a potential non-conformity that has not yet occurred, often found through trends or risk assessment. ISO 13485 keeps the three in separate clauses for a reason: a record that only shows the correction has fixed the instance and left the cause, and an inspector reads it that way. AI drafts help keep them apart by labelling what it proposes. A draft CAPA plan from an investigation carries the correction already taken, the corrective actions that address the confirmed cause, and, where the investigation found a condition that could affect other lines or products, preventive actions with their scope. The person approving the plan sees the three kinds side by side and can refuse a plan that is all correction. Three kinds of action, one record - 1 Correction Contains the instance: hold, rework, reject, replace. Done at once, recorded in the deviation. - 2 Corrective action Removes the confirmed cause: a procedure, a control, a design, a supplier. Owner, due date, evidence of implementation. - 3 Preventive action Removes a cause found by trend or risk assessment before it produces a non-conformity. Scoped to the lines and products it applies to. - 4 Effectiveness check Defined at approval: the data that will be looked at, the window, and the threshold that means the action worked. The order in which they usually happen. Only the corrective action is tested by the effectiveness check; the correction is verified by the batch disposition. ## What does AI draft in a CAPA? From the investigation's confirmed root cause and the attached documents, the AI proposes actions that address that cause, each with the role that should own it and a suggested due date, and it drafts the effectiveness check: which data to look at, over what window, and what result means success. It also reviews a CAPA plan a person wrote and says, with its reasons, whether the actions are adequate for the cause and what is missing. The plan is approved by people, under the review and approval cycle, before any action starts. During implementation the AI does not act; people complete actions and attach evidence. At verification it helps again, by reading the evidence against the criterion and drafting the effectiveness assessment for the verifier to confirm or correct. Every draft names its sources, every edit is audited with its reason, and the tool and its version are recorded on the record. Where the AI works in a CAPA, and where it does not - DataThe investigation and its evidenceConfirmed root cause, the tested hypotheses, attached SOPs and records, earlier CAPAs that share the cause. - AIDraft plan, adequacy review, effectiveness criterionActions that address the cause, with owner roles and dates; a review of a human-written plan; the data, window and threshold for the effectiveness check; a draft effectiveness assessment at verification. - People decideApprove, implement, verify, closePlan approval with signatures, implementation with evidence, verification against the criterion, closure with the effectiveness result signed. The AI drafts before approval and at verification; between the two, people implement. It never completes an action or closes a CAPA. ## How do you define an effectiveness check before closing? An effectiveness check is a measurement, decided before the action starts: the data that will show whether the cause is gone, the period over which it will be watched, and the threshold that counts as success. A check that says the SOP was updated and training was done measures the action, not its effect. A check that says no recurrence of the defect in the next twenty batches, or a defect rate below a stated level over three months, measures the effect. Writing the criterion at approval has a second benefit: it tells the team what data to keep. A control chart that was started when the CAPA was approved is the effectiveness evidence; one that was reconstructed at closure is an argument. Effectiveness checks that can be verified Root cause Corrective action Effectiveness criterion (set at approval) Worn stopper feeder rail shedding fragments Six-monthly PM made mandatory in the CMMS; rail inspection added to line clearance No particle rejects attributable to the feeder in the next 20 lots; PM completed on time for 12 months Assay method transfer missed a column lot sensitivity Method revalidated with the column lot as a parameter; specification of column lots added No OOS or OOT assay result from column lot variation in 6 months of release testing Label mix-up: two products share a storage bay Physical segregation; scanner check at line set-up Zero label mix-ups and zero scanner overrides without QA approval in 3 months Deviations closed late because of a vague escalation rule Escalation rule rewritten with named roles and days Median closure under 30 days and no unexplained extension for two quarters Examples of the form a criterion should take. The values are illustrative; each check is set against the site's own history and risk. ## How does AI detect recurrence across records? Recurrence hides in numbering. The same cause appears as a deviation one month, a complaint the next and a batch review exception after that, each with its own record and investigator. A quality system that connects records by the identifiers they share, the lot and batch numbers, the equipment, the product, the document numbers and the wording of the root cause, shows the CAPA owner the earlier records with the same cause when the plan is written, and shows the deviation investigator the open CAPA when a new event lands. Two other tools close the loop. Signal detection trends batch and laboratory results with control chart rules, so a sustained shift is flagged before it becomes a deviation and the effectiveness check has data from the day the CAPA was approved. Backlog triage ranks the open deviations, CAPAs and change controls by risk, so the CAPA that is late and matters is at the top of the list rather than the one that is late and does not. The CAPA loop - 1 . InvestigationHypotheses tested, root cause confirmed with evidence. - 2 . Plan approvalActions, owners, due dates and the effectiveness criterion, approved and signed. - 3 . ImplementationActions completed with evidence attached. - 4 . VerificationEach action checked as implemented and not harmful. - 5 . Effectiveness checkThe criterion measured over its window. - 6 . Closure and monitoringClosed with the result signed; trending continues and feeds management review. - The active step advances to show the sequence; every step is listed in the text. Each step hands to the next, and monitoring feeds back into new investigations when the data says the cause is back. The loop is the quality system's learning mechanism; a CAPA closed without the check breaks it. ## What stays with people - Approval of the plan, with electronic signatures, and the decision that the actions fit the cause. - Implementation and the evidence of it. - Verification that each action was done and did no harm, and the effectiveness assessment against the criterion. - Closure, and the escalation when the check fails: a reopened investigation, not a quietly extended date. - The assurance of the AI tool for its intended use, and change control over the model that drafts. ## Common questions ### What is the difference between a correction and a corrective action? A correction fixes the non-conformity in front of you: the batch on hold, the rework, the replaced part. A corrective action removes the cause so the non-conformity does not happen again. A preventive action removes the cause of a potential non-conformity that has not yet occurred. ISO 13485, as incorporated by the QMSR, treats corrective and preventive action in separate clauses, and the record should keep the three apart. ### Who approves a CAPA plan? The plan is approved by qualified people under the firm's procedure, usually the quality unit with the owning department, before any action starts. In QLogix this is a review and approval cycle with electronic signatures; the AI's draft plan is an input to that approval, not a substitute for it. ### Can AI close a CAPA? No. Closure needs evidence that the actions were implemented, verified and effective against the criterion set at approval, and a signature from the person responsible. AI can draft the effectiveness assessment from the evidence for the verifier to confirm, which saves time without moving the decision. ### How long should a CAPA stay open before the effectiveness check? As long as the criterion needs: twenty batches, three months of complaint data, two PM cycles. The window is set when the plan is approved, from how often the cause had the chance to recur. A CAPA closed before its window has elapsed has not been checked, whatever the record says. ## Sources - [FDA: Q10 Pharmaceutical Quality System, guidance for industry](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q10-pharmaceutical-quality-system) - [21 CFR 211.192, Production record review (eCFR, current)](https://www.ecfr.gov/current/title-21/section-211.192) - [EudraLex Volume 4, Chapter 1: Pharmaceutical Quality System (2013)](https://health.ec.europa.eu/system/files/2016-11/vol4-chap1_2013-01_en_0.pdf) - [FDA: Quality Management System Regulation (QMSR)](https://www.fda.gov/medical-devices/postmarket-requirements-devices/quality-management-system-regulation-qmsr) - [Federal Register, 2 February 2024: Medical Devices; Quality System Regulation Amendments (QMSR final rule, effective 2 February 2026)](https://www.federalregister.gov/documents/2024/02/02/2024-01709/medical-devices-quality-system-regulation-amendments) - [21 CFR Part 820, Quality Management System Regulation (eCFR, current)](https://www.ecfr.gov/current/title-21/part-820) - [FDA: Q9(R1) Quality Risk Management, guidance for industry](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q9r1-quality-risk-management) - [FDA: Inspection Observations (annual observation data)](https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/inspection-references/inspection-observations) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - CAPA - ICH Q10 - QMSR - Effectiveness checks ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [Connected quality data](https://qlogix.ai/services/connected-quality-data) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI in document control and validation workflows | QLogix Source: https://qlogix.ai/blog/ai-document-management-validation-workflow Validation # AI in document management and validation workflows for life sciences By QLogix · Published 14 October 2026 · 7 min read The short answer Document control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 is about controlled versions, approvals, effective dates, distribution and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. The same holds for validation: plans, protocols and reports drafted from templates and requirements, with assurance effort set by risk under GAMP 5 and FDA's CSA guidance, and a person approving every version with an electronic signature. EU GMP Chapter 4 (documentation) and Annex 11 (computerised systems) revised; both still in force. 2011 GAMP 5 Second Edition: critical thinking and risk-based assurance over scripting everything. 2022 FDA's computer software assurance guidance final: assurance effort set by process risk. Sept 2025 ## Key takeaways - Document control under EU GMP Chapter 4, 21 CFR Part 11 and Annex 11 means controlled versions, approvals by authorised people, effective dates, controlled distribution, training and periodic review. AI changes who writes the first draft and who finds the gaps, not what must be controlled. - The useful AI jobs in a document system are drafting from a template and the approved sources, reviewing a draft against procedures and regulations with findings by severity, and flagging what is due for periodic review or affected by a change. - Validation benefits the same way: validation plans with activities and owners, protocols and reports generated from Part 11-ready templates, test cases drafted from requirements with traceability, and assurance effort set by risk under GAMP 5 Second Edition and FDA's CSA guidance. - Everything an AI writes stays a draft until a qualified person approves it with an electronic signature. The audit trail records the tool and its version; a change of model goes through change control; exported copies are stamped and locked. ## What must a document management system control? EU GMP Chapter 4 sets the expectations most sites work to: documents are designed, prepared, reviewed and distributed with care; approved, signed and dated by appropriate and authorised people; have unambiguous contents and a unique identifier; are regularly reviewed and kept current; and superseded versions are prevented from unintended use. Annex 11 adds the controls for the computerised systems that hold them: validation, audit trails, access control, electronic signatures with the same impact as handwritten ones, and data integrity through the record's life. In the United States, 21 CFR Part 11 sets the conditions under which electronic records and signatures are trustworthy and equivalent to paper: validated systems, accurate and complete copies, record protection and retrieval, limited access, audit trails that are computer-generated and time-stamped, and signatures linked to their records. FDA's 2003 scope guidance narrowed enforcement to the records that predicate rules require, which is where the quality system's documents sit. For devices, the QMSR incorporates ISO 13485:2016, whose document control clause asks for the same things in its own words. Document control requirements and where AI can help Requirement Source Where AI helps Who decides Unique identifier, controlled versions Chapter 4; ISO 13485 Numbering by document type; a version is created, never overwritten The system enforces; the author requests Review and approval by authorised people Chapter 4; Part 11 Draft from template and sources; review findings by severity before routing Reviewers and approvers, with electronic signatures Regular review, kept current Chapter 4 Periodic review due dates; the documents a change control or regulation update affects Document owner Controlled distribution; superseded versions withdrawn Chapter 4 Stamped controlled copies; superseded versions marked for reference only Document control Training before use Chapter 2 and 4; ISO 13485 Who must attest; reminders; the attestation signed per version Each trained person; the supervisor Audit trail, access control, signatures Annex 11; Part 11 Nothing: these are system controls the AI runs inside Quality unit and IT Our reading of Chapter 4, Annex 11 and Part 11. The requirement column is the regulation's; the AI column is what a drafting and review tool can do without taking the decision. ## Where does AI fit in the document lifecycle? At the start, when a document is drafted from a template and the approved sources that constrain it: the parent procedure, the regulation, the equipment manual, the previous version. At review, where the AI reads the draft against those sources and the firm's conventions and returns findings with a severity and a location, so the author fixes them before the reviewers see the draft. At change, where a change control names the documents it affects and the AI proposes the edits for the owner to accept. And at periodic review, where the AI summarises what has changed in the regulations and in the connected records since the document was last approved. What it does not do is approve, make effective, distribute or train. Those steps carry electronic signatures and the audit trail, and they stay with the people the procedure names. The document lifecycle with AI in the loop - 1 . DraftFrom a template and the approved sources, by the author with the AI's draft. - 2 . AI reviewFindings by severity against procedures, regulations and conventions; the author resolves them. - 3 . Review and approvalReviewers and approvers sign electronically; the version is locked. - 4 . Effective dateThe version becomes the one in use; the previous one is superseded. - 5 . TrainingEach person who must follow it attests, per version, with a signature. - 6 . Periodic reviewOn the schedule, or when a change or regulation touches it; the AI summarises what changed. - The active step advances to show the sequence; every step is listed in the text. The AI contributes at drafting, review and periodic review; the signed steps are people's. Every version is kept, and the effective version is the only one in use. ## What does an AI document review look like? - Findings, each with a location in the document, a category and a severity: major, minor or observation. - Clarity: steps that can be read two ways, undefined terms, instructions without an actor. - Completeness: a section the template expects that is missing, a form referenced but not attached, a record the step produces that is not named. - Consistency: a limit that differs from the specification, a reference to a superseded version, a responsibility that contradicts the parent procedure. - Compliance: a requirement from the regulation or the standard the document should implement that it does not address, cited by clause. - Data integrity: a step that records a result without saying where, how and by whom, or that allows a value to be overwritten. - Safety: a step with a hazard and no control. Where review findings fall in a first AI review - Clarityambiguous steps, undefined terms34 % - Consistencylimits, references, responsibilities24 % - Completenessmissing sections or records19 % - Complianceclause not addressed11 % - Data integritywhere, how, by whom8 % - Safetyhazard without a control4 % An illustrative distribution for a procedure drafted from a template, from our experience of such reviews, not a published statistic. Clarity and consistency findings dominate; compliance findings are fewer but carry the most weight. ## How do controlled copies and training stay controlled with AI in the loop? A controlled copy is a copy that says what it is. The approved version carries its stamp with the final approval and effective dates and the approvers; a draft says draft and is not for use; a version routed for approval says so; a superseded version is for reference only. Printed copies are uncontrolled by definition, and the stamp says that too. The copy is locked so it cannot be edited after it leaves the system, and the audit trail records who exported which version when. Training follows the version, not the document: when a new version becomes effective, the people who must follow it attest that they have read and understood that version, with an electronic signature, and the record shows who has and who has not. The AI's part is small and useful: proposing who is affected from the roles the document names, and summarising what changed between versions so the attestation is informed. ## How does AI change validation workflow? Validation work is document-heavy by design: a plan, requirements, risk assessments, protocols, executed tests, deviations, reports. Each has a template, a set of inputs and an approval, which is the shape AI drafting suits. From a description of the system and the attached documents, the AI drafts the validation plan with its activities, owners and due dates; from a template and the requirements, it drafts a protocol with test cases traced to requirements; from the executed results and deviations, it drafts the summary report. Each is a draft for the validation lead to correct and the approvers to sign, and a protocol is approved before it is executed and a report after. GAMP 5 Second Edition and FDA's computer software assurance guidance set the effort: critical thinking about what the system does and what happens if it fails, scripted testing where the risk calls for repeatability and traceability, unscripted testing where it does not, and supplier evidence where an assessment supports relying on it. A drafting tool does not change that calculus; it makes the documents that record it faster to produce and more consistent. A computerised system validation, requirement by requirement - User requirementsURS, traced120 - Assessed high process riskGxP impact, data integrity, safety34 - Scripted tests (high risk)steps, expected, actual34 - Unscripted tests (the rest)exploratory, with a record86 - Executed with results recordedpass, fail, deviation120 - Deviations raised and closedbefore the report7 - Requirements verified in the summary reporttraceability complete120 An illustrative model of one system's validation under a risk-based approach, not a measurement. Requirements flow to tests whose rigor follows risk; the counts show that most requirements do not need scripted testing. Each stage is a count of requirements. ## Which assurance effort does the AI tool itself need? The same question applies to the AI feature as to any software in the quality system: what does it do, and what happens if it is wrong? A feature that drafts a document or a protocol for a person to review sits at a lower process risk than one that changes a record or releases a batch, and the assurance activities follow. Two things are specific to AI. Its behaviour can change when the model or the prompt changes, so the assurance record names the model version and a change goes through change control with regression tests. And the draft must never be mistaken for the record: it is a version, audited like any other, until a person approves it. In the European Union, a draft Annex 22 on artificial intelligence in GMP was put to consultation in 2025 and had not been finalised when this article was last updated. Its direction is consistent with the above: intended use defined, performance evidenced, a human in the loop where the output affects quality decisions, and the model under change control. Assuring an AI drafting feature - 1 Intended use What the feature drafts, from which inputs, for whom to review. What it must not do. - 2 Process risk What happens if the draft is wrong and the review misses it; the controls that catch it. - 3 Assurance activities Exploratory testing of drafts against known inputs; scripted tests of the controls (audit trail, signature, version). - 4 Model and prompt under control Versions named in the record; changes through change control with regression tests. - 5 Periodic review Performance and incidents reviewed on a schedule, with the documents the feature drafts. The sequence we follow under GAMP 5 Second Edition and the CSA guidance. The record is proportionate to the process risk of the feature, not to the novelty of the technology. ## What stays with people - Approval of every version, with electronic signatures, and the decision to make it effective. - The review findings' dispositions: accepted, rejected with a reason, deferred to a change. - Training attestations and the decision of who must train. - Protocol approval before execution, execution itself, and the report's approval after. - The risk assessment of the AI tool, the change control over its model, and the periodic review of its performance. ## Common questions ### Can an AI-drafted SOP be approved as it is? Only by the people the procedure names, after review, with electronic signatures, exactly as a human-drafted SOP. The AI's draft is a version of the document with its author recorded; the reviewers' findings and the approvers' signatures make it an approved document. Nothing about the drafting changes the approval. ### Does 21 CFR Part 11 apply to AI-generated documents? Part 11 applies to the electronic records and signatures that predicate rules require, however the text was produced. The system that holds an AI-drafted procedure must be validated, keep an audit trail, control access, and link signatures to their records. The AI's contribution is recorded in that audit trail with the tool and its version. ### Do AI document and validation tools need to be validated? They need assurance for their intended use, proportionate to process risk, under GAMP 5 Second Edition and FDA's CSA guidance. A drafting feature whose output a person reviews is lower risk than a feature that changes a record. The assurance record names the model version, and a change of model goes through change control with regression testing. ### What is the difference between CSA and GAMP 5? FDA's computer software assurance guidance, final since September 2025, is the agency's statement of a risk-based approach for software used in medical device production and quality systems. GAMP 5 Second Edition, published by ISPE in 2022, is the industry's good practice guide for compliant GxP computerised systems across drugs and devices. They point the same way: critical thinking, assurance effort by risk, and records that show the reasoning rather than screenshots of every step. ## Sources - [EudraLex Volume 4, Chapter 4: Documentation (2011)](https://health.ec.europa.eu/system/files/2016-11/chapter4_01-2011_en_0.pdf) - [EudraLex Volume 4, Annex 11: Computerised Systems (2011)](https://health.ec.europa.eu/system/files/2016-11/annex11_01-2011_en_0.pdf) - [FDA: Part 11, Electronic Records; Electronic Signatures: Scope and Application, guidance for industry (2003)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/part-11-electronic-records-electronic-signatures-scope-and-application) - [ISPE: GAMP 5 Guide, Second Edition (2022)](https://ispe.org/publications/guidance-documents/gamp-5-guide-2nd-edition) - [Federal Register, 24 September 2025: Computer Software Assurance for Production and Quality System Software, final guidance, notice of availability](https://www.federalregister.gov/documents/2025/09/24/2025-18468/computer-software-assurance-for-production-and-quality-system-software-guidance-for-industry-and) - [FDA: Quality Management System Regulation (QMSR)](https://www.fda.gov/medical-devices/postmarket-requirements-devices/quality-management-system-regulation-qmsr) - [European Commission: stakeholder consultation on EudraLex Volume 4, Chapter 4 and Annex 22 (2025)](https://health.ec.europa.eu/consultations/stakeholders-consultation-eudralex-volume-4-good-manufacturing-practice-guidelines-chapter-4-annex_en) - [MHRA: GxP Data Integrity Guidance and Definitions (2018)](https://www.gov.uk/government/publications/guidance-on-gxp-data-integrity) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Document control - Part 11 - Annex 11 - GAMP 5 - CSA ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # FDA's 7-step AI credibility framework explained | QLogix Source: https://qlogix.ai/blog/fda-ai-credibility-framework AI in quality # FDA's seven-step credibility framework for AI, in plain terms By QLogix · Published 12 October 2026 · 4 min read The short answer FDA's draft guidance asks sponsors to show that an AI model is credible for a specific context of use. The effort scales with model risk: how much the model's output drives the decision, and how serious a wrong decision would be. Steps in the credibility assessment framework. 7 Drug and biologic submissions with AI components FDA has seen since 2016. 500+ Factors that set model risk: model influence and decision consequence. 2 ## Key takeaways - FDA published the draft guidance in January 2025. It was still a draft when this article was last updated. - It applies when an AI model produces information used to support a regulatory decision about the safety, effectiveness or quality of a drug or biological product, including in manufacturing. - Credibility is trust in the model's performance for one context of use, supported by evidence. It is not a property of the model in general. - Model risk combines two things: how much the output influences the decision, and the consequence if the decision is wrong. ## What is the FDA draft guidance on AI for drugs and biologics? In January 2025 FDA published the draft guidance Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products. It gives sponsors a risk-based framework for establishing and assessing the credibility of an AI model for a particular context of use. The comment period closed in April 2025. A draft guidance describes FDA's current thinking and is not binding. It is still the clearest statement of what the agency expects to see when AI output supports a decision it will review. ## When does it apply? It applies to AI models that produce information or data intended to support regulatory decisions about the safety, effectiveness or quality of drugs and biological products. That reaches across the product life cycle: nonclinical work, clinical trials, postmarketing surveillance and manufacturing. It does not address AI used in drug discovery, or AI used for operational efficiency, such as drafting or resource planning, where the use does not affect patient safety, drug quality or the reliability of study results. ## What are the seven steps? - Step 1. Define the question of interest: the specific question or decision the model will address. - Step 2. Define the context of use: what the model does, what it is used for, and whether other evidence is used alongside its output. - Step 3. Assess the AI model risk, from model influence and decision consequence. - Step 4. Develop a plan to establish the credibility of the model output within the context of use. - Step 5. Execute the plan. - Step 6. Document the results of the credibility assessment plan and discuss deviations from it. - Step 7. Determine whether the model is adequate for the context of use. The seven-step credibility assessment - 1 Question of interest The decision the model informs - 2 Context of use The model's role and scope - 3 Model risk Influence and consequence - 4 Credibility plan How the output will be shown to be trustworthy - 5 Execute Run the plan - 6 Document Results and deviations - 7 Adequacy Fit for the context of use, or not Steps 1 to 3 decide how much evidence is needed. Steps 4 to 7 produce it and judge it. ## How is AI model risk assessed? Model risk is not the risk of the model in the abstract. It is the risk that its output leads to a wrong decision with an adverse outcome. The draft builds it from two factors. - Model influence: how much the AI output contributes to the decision, compared with other evidence. A model whose output is confirmed by an independent test has less influence than one that decides alone. - Decision consequence: how serious the outcome is if the decision turns out to be wrong. How the two factors combine into model risk Model influence Low Medium High High Medium High High Medium Low Medium High Low Low Low Medium Decision consequence An illustration of the draft guidance's two factors. The draft does not prescribe a scoring grid; sponsors justify the risk level for their own context of use. ## What does this look like in manufacturing and quality? Take a model that screens batch data and flags batches for closer review. If every batch still goes through the full release testing and a qualified person makes the release decision, the model's influence is low: it directs attention and decides nothing. If the same model were used to skip testing on batches it scores as normal, its influence would be high, and so would the evidence FDA expects. The practical lesson is that design choices move the risk. Keeping a qualified reviewer and an independent check in the process lowers model influence, and that lowers the validation burden along with the risk to the patient. ## What should you document? - The question of interest and the context of use, written precisely enough that a reviewer can tell what is in and out of scope. - The model and how it was developed: the data used for training and tuning, and how it was managed. - How the model was evaluated: test data independent of development data, the performance metrics and the acceptance criteria. - Life cycle maintenance: how performance is monitored after deployment and how changes to the model are controlled. ## Common questions ### Is the FDA AI guidance for drugs and biologics final? It was still a draft when this article was last updated in October 2026. FDA published it in January 2025 and the comment period closed in April 2025. ### What is a context of use? The context of use defines the specific role and scope of the AI model used to address a question of interest: what is modelled, how the output is used, and whether other information is used with it to answer the question. ### Does the FDA draft guidance cover AI that drafts documents? Not where the use is only for operational efficiency and does not affect patient safety, drug quality or the reliability of study results. The draft sets those uses outside its scope. Your own quality system still decides how such tools are controlled. ## Sources - [FDA draft guidance: Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products (January 2025)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/considerations-use-artificial-intelligence-support-regulatory-decision-making-drug-and-biological) - [FDA press announcement, 6 January 2025: FDA proposes framework to advance credibility of AI models used for drug and biological product submissions](https://www.fda.gov/news-events/press-announcements/fda-proposes-framework-advance-credibility-ai-models-used-drug-and-biological-product-submissions) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - FDA - AI credibility - Context of use - Drug manufacturing ## How QLogix helps - [Connected quality data](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI in drug manufacturing: where FDA stands | QLogix Source: https://qlogix.ai/blog/ai-in-drug-manufacturing-fda Manufacturing science # AI in drug manufacturing: where FDA stands and what to build for By QLogix · Published 12 October 2026 · 4 min read The short answer FDA has no final guidance written only for AI in drug manufacturing. It has a 2023 discussion paper naming five open areas and a 2025 draft framework that covers manufacturing models. Together they show what a process model has to prove. Areas of consideration in FDA's 2023 discussion paper on AI in drug manufacturing. 5 Year of FDA's PAT guidance, the starting point for model-based control. 2004 Draft credibility guidance, which covers manufacturing models. Jan 2025 ## Key takeaways - FDA's drug center published the discussion paper Artificial Intelligence in Drug Manufacturing on 1 March 2023. It asks questions; it sets no requirements. - The January 2025 draft guidance on AI credibility applies to models that support decisions about drug quality, which includes models used in manufacturing. - A model that advises an operator carries less regulatory weight than one that controls the process or replaces a release test. - Existing CGMP rules already apply: equipment and automated systems under 21 CFR 211.68, in-process controls under 211.110, and change control for any model update. ## Does FDA have guidance on AI in drug manufacturing? Not a final guidance written only for that subject. What exists is a line of documents that build on one another. The 2004 guidance on process analytical technology (PAT) accepted real-time measurement and model-based control as part of pharmaceutical quality. ICH Q13, published by FDA as final guidance in March 2023, set expectations for continuous manufacturing, where process models and automated control do most of the work. In March 2023 FDA's Center for Drug Evaluation and Research published a discussion paper on AI in drug manufacturing and asked for comment, followed by a public workshop with PQRI in September 2023. In January 2025 the agency published its draft guidance on establishing the credibility of AI models used to support regulatory decisions for drugs and biological products. That draft applies across the product life cycle, manufacturing included. FDA documents that shape AI in drug manufacturing - Sep 2004Final PAT framework guidance Real-time measurement and model-based control accepted as part of quality assurance. - Mar 2023Final ICH Q13, continuous manufacturing Expectations for process models, control strategy and batch definition in continuous processes. - Mar 2023Discussion paper Discussion paper: AI in drug manufacturing Five areas of consideration, published for comment. - Sep 2023Workshop FDA and PQRI workshop Public workshop on the regulatory framework for AI in manufacturing. - Jan 2025Draft Draft guidance: AI to support regulatory decision-making Seven-step credibility framework, covering manufacturing models. Dates of publication by FDA. Status as of October 2026. ## What are the five areas in FDA's discussion paper? The paper covers drug products marketed under an NDA, ANDA or BLA. It names five areas where the current framework raises questions, and states that the list is not exhaustive. - Cloud applications may affect oversight of pharmaceutical manufacturing data and records. - The Internet of Things may increase the amount of data generated during manufacturing, which affects existing data management practices. - Applicants may need clarity about whether and how the application of AI in manufacturing is subject to regulatory oversight. - Applicants may need standards for developing and validating AI models used for process control and to support release testing. - Continuously learning AI systems that adapt to real-time data may challenge regulatory assessment and oversight. ## Where does AI sit in a manufacturing control architecture? The regulatory weight of a model depends on where its output goes. The same soft sensor can be a monitoring aid, an input to a control loop or the basis for releasing a batch without an end-product test. Each step up moves decisions from a person to the model, and raises what the model has to prove. This is the same idea as model influence in the 2025 draft guidance. A design that keeps an independent measurement and a qualified person in the release decision keeps the model's influence low. From sensor to release decision - People decideBatch dispositionQuality unit reviews the record and releases or rejects the batch. - People decideReview of alarms and excursionsOperators and QA assess model alerts, with the reason recorded. - AIProcess modelSoft sensor, multivariate model or digital twin predicts a quality attribute and flags drift. - DataData platformContextualized, time-aligned process and laboratory data with audit trails. - DataSensors and analyzersPAT instruments, equipment signals, in-process tests. A reference architecture for a model-based process control. The orange layer is where AI works; the outlined layers are decisions kept with qualified people. ## What does a manufacturing model have to show? - A stated context of use: which attribute it predicts, on which product and equipment, and what the output is used for. - Development data that covers the process range the model will see, including the variation between lots, materials and seasons. - Performance on data that played no part in building it, against acceptance criteria set in advance. - Behaviour outside its range: what happens when an input drifts or a sensor fails, and how the system falls back to a conventional control. - Monitoring in routine use, with triggers for review and retraining. - Change control for every model update. FDA's paper notes that deployed models are updated as needed through the change control process, which is the practical limit on continuous learning today. ## How should a manufacturing science team start? - Pick a use where the model advises and a person decides, such as predicting a drift before it reaches an action limit. - Get the data foundation right first. A model trained on data without context or audit trails cannot be defended in an inspection. - Write the context of use before building, and size the validation to the influence the model will have. - Talk to FDA early for uses that touch the control strategy or release testing. The Emerging Technology Program exists for that conversation. ## Common questions ### Is there a final FDA guidance on AI in drug manufacturing? No. As of October 2026 FDA has a 2023 discussion paper on AI in drug manufacturing and a January 2025 draft guidance on AI credibility that covers manufacturing models. Existing CGMP regulations and the PAT and ICH Q13 guidances continue to apply. ### Can a continuously learning model be used in GMP manufacturing? Not as a model that changes by itself in production. FDA's discussion paper lists continuously learning systems as an open challenge, and in practice models are locked and updated through change control with revalidation. ### Does a digital twin need to be validated? Yes, when its output supports a GMP decision. The depth of validation follows its use: a twin used to explore scenarios offline needs less than one that feeds a control loop or supports batch release. ## Sources - [FDA CDER discussion paper: Artificial Intelligence in Drug Manufacturing (March 2023)](https://www.fda.gov/media/165743/download) - [FDA draft guidance: Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products (January 2025)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/considerations-use-artificial-intelligence-support-regulatory-decision-making-drug-and-biological) - [FDA guidance: Q13 Continuous Manufacturing of Drug Substances and Drug Products (March 2023)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/q13-continuous-manufacturing-drug-substances-and-drug-products) - [FDA guidance: PAT, A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance (September 2004)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/pat-framework-innovative-pharmaceutical-development-manufacturing-and-quality-assurance) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Drug manufacturing - Process analytical technology - Continuous manufacturing - Process models - FDA ## How QLogix helps - [Connected quality data](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # AI-enabled medical devices: FDA list and guidance | QLogix Source: https://qlogix.ai/blog/ai-enabled-medical-devices-fda Medical devices # AI-enabled medical devices: what FDA's list and guidance show By QLogix · Published 12 October 2026 · 4 min read The short answer FDA's public list counted 1,614 AI-enabled medical devices through June 2026, about three in four of them in radiology. Two documents shape how they are built and changed: the final guidance on predetermined change control plans and the draft guidance on lifecycle management. AI-enabled devices on FDA's list through June 2026. 1,614 Share of those authorizations in radiology (1,230 devices). 76% AI-enabled devices authorized in the second quarter of 2026. 89 ## Key takeaways - FDA's list of AI-enabled medical devices reached 1,614 through June 2026, according to an analysis of the list by The Imaging Wire. FDA says the list is not comprehensive. - Radiology accounts for about three in four authorizations, a share that has stayed between 73% and 80% a year since 2023. - A predetermined change control plan lets a manufacturer make the model changes described in the plan without a new marketing submission for each one. - The draft lifecycle guidance of January 2025 sets out what a marketing submission for an AI-enabled device should contain. It was still a draft when this article was last updated. ## How many AI-enabled medical devices has FDA authorized? FDA publishes a list of AI-enabled medical devices that have been authorized for marketing in the United States and updates it periodically. An analysis of the list published in September 2026 counted 1,614 devices through the end of June 2026, up 5.9% on the first quarter. FDA authorized 92 AI-enabled devices in the first quarter of 2026 and 89 in the second. FDA describes the list as a resource for the public and not a comprehensive record. Devices are identified from AI-related terms in the public authorization summaries, so a device that uses AI without saying so in its summary may be missing. ## Which specialties do they serve? Radiology dominates. Of the 1,614 devices, 1,230 are radiology devices, which is 76%. Image analysis suits machine learning: the input is digital, large labelled datasets exist, and the output can be checked against a reader. Cardiovascular and neurology devices follow at a distance. Radiology's share of FDA's AI-enabled device authorizations - 202380 % - 202473 % - 202575 % - Q1 202692 devices authorized75 % - Q2 202689 devices authorized74 % Share of AI-enabled device authorizations that were radiology devices, by period. Source: The Imaging Wire analysis of FDA's AI-enabled medical device list, September 2026. ## What is a predetermined change control plan? A predetermined change control plan, or PCCP, is a part of a marketing submission in which the manufacturer describes the modifications it intends to make to an AI-enabled device software function after authorization, and how it will make them. Congress gave FDA the authority to accept such plans in 2022, in section 515C of the Federal Food, Drug, and Cosmetic Act. FDA's guidance on PCCPs for AI-enabled device software functions is final; it was first issued in December 2024 and the current version is dated August 2025. Once FDA has authorized a device with a PCCP, the manufacturer can implement the changes the plan covers without a new submission for each one, provided it follows the plan. The three parts of a PCCP - 1 Description of modifications The specific, planned changes to the model - 2 Modification protocol Data management, retraining, performance evaluation and update procedures - 3 Impact assessment Benefits and risks of the planned changes, and how the risks are mitigated The components FDA's final guidance recommends for a predetermined change control plan. ## What does the draft lifecycle guidance ask for? In January 2025 FDA published the draft guidance Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations. It takes a total product life cycle view and lists what a marketing submission should contain. - A device description that says what the AI does, its inputs and outputs, and who uses it. - User interface and labelling that tell the user what the model was trained on, how it performs and where its limits are. - A risk assessment that covers AI-specific risks, including bias and performance that varies between patient groups. - Data management: how training and test data were collected, kept separate and shown to represent the intended population. - Model description, development and validation, with performance broken down by relevant subgroups. - A plan for monitoring performance after the device is on the market. ## Where do robotics fit in? Robotically-assisted surgical devices are regulated as medical devices in their own right, and a growing number add AI features such as image guidance and motion planning. When they do, the same expectations apply to the AI function: a defined intended use, evidence of performance, transparency to the user and a plan for changes. On the manufacturing side, device makers now work under the Quality Management System Regulation, which took effect on 2 February 2026 and incorporates ISO 13485:2016. Software used in production and in the quality system, AI-enabled or not, is validated for its intended use with effort proportionate to risk. ## What should a device quality team do now? - Decide early whether the model will change after launch. If it will, plan the PCCP with the first submission instead of retrofitting it. - Treat training and test data as controlled records, with provenance, separation and representativeness documented. - Build post-market performance monitoring into the quality system, with complaint handling and CAPA connected to it. - Apply the ten Good Machine Learning Practice guiding principles that FDA, Health Canada and the MHRA published in 2021 as a design checklist. ## Common questions ### How many AI-enabled medical devices has FDA authorized? FDA's public list counted 1,614 AI-enabled medical devices through June 2026, according to an analysis published by The Imaging Wire in September 2026. About 76% are radiology devices. FDA notes the list is not comprehensive. ### Is the PCCP guidance for AI-enabled devices final? Yes. FDA's guidance on marketing submission recommendations for a predetermined change control plan for AI-enabled device software functions is final. It was first issued in December 2024 and the current version is dated August 2025. ### Can an AI-enabled device learn continuously after it is authorized? Changes to the model need to be covered by an authorized predetermined change control plan or by a new marketing submission where one is required. A PCCP describes the planned changes, the protocol for making them and their impact in advance. ## Sources - [FDA: Artificial Intelligence-Enabled Medical Devices (the list)](https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-enabled-medical-devices) - [The Imaging Wire, 13 September 2026: analysis of FDA's AI-enabled device list through June 2026](https://theimagingwire.com/2026/09/13/radiology-maintained-its-lead-in-fda-approvals/) - [FDA final guidance: Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial-intelligence) - [FDA draft guidance: Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations (January 2025)](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/artificial-intelligence-enabled-device-software-functions-lifecycle-management-and-marketing) - [FDA: Quality Management System Regulation (QMSR)](https://www.fda.gov/medical-devices/postmarket-requirements-devices/quality-management-system-regulation-qmsr) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - AI-enabled devices - PCCP - SaMD - Radiology AI - Surgical robotics - QMSR ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Robotics in aseptic manufacturing and EU GMP Annex 1 | QLogix Source: https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1 Manufacturing science # Robotics in aseptic manufacturing: what Annex 1 expects By QLogix · Published 12 October 2026 · 3 min read The short answer The revised EU GMP Annex 1 names robotic systems, with isolators and RABS, as technologies to consider for protecting sterile products. Taking the operator out of the critical zone removes the largest contamination source and moves the quality work to design, software and data. Revised EU GMP Annex 1 came into operation. 25 Aug 2023 Where Annex 1 names robotic systems among technologies to consider. Section 2.1 Gloves on a gloveless robotic isolator: no direct human intervention in the critical zone. 0 ## Key takeaways - Annex 1 does not require robots. It says appropriate technologies, including RABS, isolators and robotic systems, should be considered to increase the protection of the product. - People are the main source of microbial contamination in aseptic processing, which is why removing interventions is the strongest control available. - A robotic line still needs a contamination control strategy, environmental monitoring and aseptic process simulation. - The robot's software and recipes are GMP computerised systems, validated and kept under change control. ## What does Annex 1 say about robotics? The revised Annex 1, Manufacture of Sterile Medicinal Products, was published in August 2022 and came into operation on 25 August 2023, with one provision on lyophilizer loading following a year later. Section 2.1 says that the use of appropriate technologies, giving restricted access barrier systems (RABS), isolators, robotic systems, rapid or alternative methods and continuous monitoring systems as examples, should be considered to increase the protection of the product from potential extraneous sources of contamination. The annex treats barrier technology as the expected direction. It describes RABS and isolators as beneficial in assuring the required conditions and minimizing microbial contamination from direct human interventions in the critical zone, and asks that alternative approaches be justified. Annex 1 revision dates - Aug 2022Published Revised Annex 1 published Adds the contamination control strategy and quality risk management throughout. - 25 Aug 2023In force Comes into operation Applies to the manufacture of sterile medicinal products. - 25 Aug 2024In force Point 8.123 applies The provision on lyophilizer loading and unloading, given an extra year. Dates from the European Commission's publication of the revised annex. ## How do barrier and robotic systems compare? Each step along this range reduces how often a person reaches into the zone where sterile product is exposed. A gloveless robotic isolator removes the glove ports altogether: format parts, components and filled units are handled by the robot, and the enclosure stays closed from decontamination to the end of the batch. Operator access to the critical zone, by technology Technology Operator access during the batch Main contamination control What quality has to watch Conventional cleanroom filling Gowned operators beside the line Gowning, aseptic technique, airflow Intervention counts, operator qualification RABS Through glove ports; doors closed in operation Physical barrier and airflow Glove integrity, door-open events Isolator Through glove ports only Closed enclosure with automated bio-decontamination Decontamination cycle, leak and glove testing Gloveless robotic isolator None Closed enclosure, robotic handling Robot recipes, software changes, fault recovery A general comparison. The classification and controls for a specific line come from its own contamination control strategy. ## What changes for contamination control? - The contamination control strategy has to cover the robot itself: surfaces that can be cleaned and decontaminated, materials compatible with vaporized hydrogen peroxide, and particles generated by its movement. - Robot paths are designed so the arm does not pass over open containers and disturb first air. - Environmental monitoring continues. With no operator present, sampling plates and probes have to be placed and removed by the robot or by design. - Aseptic process simulation covers the routine robot operations and the interventions that remain, such as fault recovery. - Fault handling is defined in advance: which stoppages the robot recovers from, which require the enclosure to be opened, and what happens to the product exposed at that moment. ## How are the robot and its software validated? A filling robot is equipment and a computerised system at once. Its qualification covers the mechanical side, and its software, recipes and vision systems fall under Annex 11 in the EU and the equivalent expectations elsewhere. A risk-based approach fits well: a recipe parameter that sets fill volume or stopper placement is high risk and gets scripted testing, while a screen that only displays status does not. Where machine vision or machine learning is added, for example to inspect containers or guide the arm, the model needs its own evidence: a defined intended use, test data kept apart from training data, and control over changes. ## What does the quality unit decide? - Batch disposition stays with the quality unit. A robotic line produces more data per batch and the same decision. - Excursions and alarms are reviewed by people, with the assessment recorded. - Changes to recipes, paths and software go through change control with an impact assessment. - The data the line generates, from particle counts to torque curves, is reviewed for trends, which is where early signals of drift appear. ## Common questions ### Does EU GMP Annex 1 require robotics? No. Section 2.1 says appropriate technologies such as RABS, isolators and robotic systems should be considered to increase the protection of the product. The annex expects barrier technology to be used or an alternative to be justified, and leaves the choice to the manufacturer's contamination control strategy. ### Does a robotic filling line still need media fills? Yes. Aseptic process simulation remains required. It covers the routine robotic operations and any interventions that remain, including recovery from faults. ### When did the revised Annex 1 take effect? The revised Annex 1 came into operation on 25 August 2023. Point 8.123, on lyophilizer loading and unloading, applied from 25 August 2024. ## Sources - [European Commission: EU GMP Annex 1, Manufacture of Sterile Medicinal Products (August 2022)](https://health.ec.europa.eu/system/files/2022-08/20220825_gmp-an1_en_0.pdf) - [European Commission: stakeholder consultation on EudraLex Volume 4, Chapter 4, Annex 11 and new Annex 22](https://health.ec.europa.eu/consultations/stakeholders-consultation-eudralex-volume-4-good-manufacturing-practice-guidelines-chapter-4-annex_en) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Aseptic processing - EU GMP Annex 1 - Robotics - Isolators - Contamination control strategy ## How QLogix helps - [Connected quality data](https://qlogix.ai/services/connected-quality-data) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # EU GMP Annex 22: what the draft AI annex requires | QLogix Source: https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft Regulatory updates # EU GMP Annex 22: what the draft AI annex asks of GMP manufacturers By QLogix · Published 12 October 2026 · 4 min read The short answer Draft Annex 22 is the first GMP annex written for AI. It allows static, deterministic models in critical GMP applications, keeps generative AI out of them, and asks for a defined intended use, independent test data and ongoing monitoring. Status in October 2026. No date of application has been set. Draft Close of the stakeholder consultation. 7 Oct 2025 Model type the draft accepts in critical GMP applications. Static only ## Key takeaways - Annex 22 is a draft. The consultation ran from July to 7 October 2025, and no final text had been published when this article was last updated. - The draft covers AI models used in critical GMP applications and limits them to static models with deterministic output. - Generative AI and large language models are excluded from critical GMP applications. In non-critical uses the draft expects qualified people to review the output. - The work it asks for is familiar validation work: a documented intended use, acceptance criteria set in advance, test data kept independent of training data, and monitoring in operation. ## What is EU GMP Annex 22? Annex 22, Artificial Intelligence, is a proposed new annex to EudraLex Volume 4, the EU's good manufacturing practice guidelines. The European Commission published it for consultation in July 2025 together with a revised Annex 11 (computerised systems) and a revised Chapter 4 (documentation). It was drafted jointly with PIC/S, so the same text is expected to reach PIC/S member authorities outside the EU. It is the first GMP annex written specifically for AI. Annex 11 continues to cover computerised systems in general, and Annex 22 adds requirements for the model itself: how it is selected, trained, tested and kept under control. ## Is Annex 22 in force? No. The stakeholder consultation closed on 7 October 2025. When this article was last updated, the final text had not been published and no date of application had been set. Check the European Commission's consultation page listed under Sources for the current status before you plan against a date. Draft or not, the direction is clear enough to act on. Inspectors already expect computerised systems to be validated for their intended use under Annex 11, and the draft shows how regulators intend to read that expectation for AI. ## Which AI models does the draft allow in critical GMP applications? The draft applies to AI models used in critical applications: those with a direct impact on patient safety, product quality or data integrity. For those applications it accepts only static models, meaning models that do not keep learning once they are in use, and only models with deterministic output, meaning the same input gives the same output. - Allowed in critical applications: static machine learning models with deterministic output, for example a classifier trained once and then locked. - Not to be used in critical applications: dynamic models that continue to learn in operation, and models with probabilistic output. - Not to be used in critical applications: generative AI and large language models. What draft Annex 22 allows, by model type and use Model type Critical GMP application Non-critical application Static model, deterministic output Allowed, under the annex's requirements Allowed Dynamic model that keeps learning in use Not to be used Outside the annex's scope Model with probabilistic output Not to be used Outside the annex's scope Generative AI, large language models Not to be used Allowed with qualified human review of the output Summary of the draft text published for consultation in July 2025. Critical means a direct impact on patient safety, product quality or data integrity. ## What does the draft say about generative AI? The draft states that generative AI and large language models should not be used in critical GMP applications. Where they are used in non-critical applications, it expects personnel with adequate qualification and training to be responsible for checking that the output is fit for its purpose, which is a human-in-the-loop arrangement. That is the line most quality teams already draw. A language model can draft an investigation summary, propose a root cause hypothesis or flag a record for attention. A qualified person reviews the draft, decides, and signs. The decision and the accountability stay with that person. ## What does a manufacturer have to do for a model in scope? - Describe the intended use in full, including the inputs the model is expected to see and the limits of that range. - Set acceptance criteria before testing, with performance at least as good as the process the model replaces. - Keep test data independent of the data used for training and validation, and show that it represents the intended use. - Show which features drive the model's output, and record a confidence score so that low-confidence results are flagged instead of acted on. - Keep the model under change control and configuration control, monitor its performance in operation, and watch for input data drifting outside the range it was tested on. - Where a person reviews the output, define that review in procedures and train the reviewers. ## What should a quality team do now? - List every place AI is used or planned in GMP activities, and mark each one critical or non-critical. - For generative AI, write down what the tool does, what the reviewer checks, and where the reviewer's decision is recorded. - For machine learning models in critical use, gather the intended use statement, the test data and the acceptance criteria, and identify the gaps against the draft. - Read the revised Annex 11 alongside it. Its requirements for audit trails, supplier oversight and security apply to the systems that host these models. ## Common questions ### Does Annex 22 ban generative AI in pharmaceutical manufacturing? No. The draft says generative AI and large language models should not be used in critical GMP applications, which are those with a direct impact on patient safety, product quality or data integrity. In non-critical applications they can be used with qualified people responsible for reviewing the output. ### When does Annex 22 take effect? No date has been set. The consultation closed on 7 October 2025 and the final text had not been published when this article was last updated in October 2026. ### Does Annex 22 replace Annex 11? No. Annex 22 adds requirements specific to AI models. Annex 11 continues to apply to the computerised systems those models run in, and a revised Annex 11 was published for consultation at the same time. ## Sources - [European Commission: stakeholder consultation on EudraLex Volume 4, Chapter 4, Annex 11 and new Annex 22](https://health.ec.europa.eu/consultations/stakeholders-consultation-eudralex-volume-4-good-manufacturing-practice-guidelines-chapter-4-annex_en) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - EU GMP Annex 22 - Annex 11 - AI in GMP - Human in the loop ## How QLogix helps - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Computer software assurance: FDA final guidance | QLogix Source: https://qlogix.ai/blog/computer-software-assurance-final-guidance Validation # Computer software assurance: what FDA's final guidance changes for validation By QLogix · Published 12 October 2026 · 4 min read The short answer FDA's final guidance on computer software assurance replaces test-everything validation with effort matched to risk. Features that could compromise safety get scripted testing; the rest can be covered by unscripted testing and supplier evidence. Final guidance announced in the Federal Register. 24 Sep 2025 QMSR took effect, bringing ISO 13485:2016 into 21 CFR Part 820. 2 Feb 2026 Risk levels in the guidance: high process risk, and not high process risk. 2 ## Key takeaways - FDA announced the final guidance, Computer Software Assurance for Production and Quality System Software, in the Federal Register on 24 September 2025. - It covers software used in medical device production or the quality system. It does not cover software that is itself a device or part of one. - Each software feature is judged by process risk: whether its failure could foreseeably compromise safety. - The record needs to show the intended use, the risk decision, what was tested and the conclusion. Screenshots of every step are not the measure of assurance. ## What is computer software assurance? Computer software assurance, or CSA, is a risk-based approach to establishing confidence that software used in production or in the quality system is fit for its intended use. FDA issued it as a draft in September 2022 and announced the final guidance on 24 September 2025. It is aimed at the habit that grew up around computer system validation: scripting and documenting every function to the same depth, whatever the risk. CSA asks teams to think first about what the software does and what happens if it fails, and to spend the testing effort there. ## Which software does the guidance cover? The guidance comes from FDA's device and biologics centers and applies to computers and automated data processing systems used as part of medical device production or the quality system. Examples include an electronic QMS, a manufacturing execution system, a laboratory system and the tools used to trend quality data. It does not apply to software as a medical device or software in a medical device. Drug manufacturers are not its formal audience, but many apply the same reasoning under GAMP 5 Second Edition, which takes the same risk-based, critical-thinking approach. ## How does the risk-based approach work? - Identify the intended use of the software, feature or function, and whether it is used directly in production or the quality system or only supports them. - Decide the process risk. A feature is high process risk when its failure to perform as intended may result in a quality problem that foreseeably compromises safety. - Choose assurance activities that match the risk: more rigor for high process risk, less for the rest. - Keep a record that is enough to show the software was assessed and performs as intended. ## Which testing methods does the guidance accept? - Unscripted testing: ad hoc testing, error guessing and exploratory testing, where the tester works from the objective without a step-by-step script. - Scripted testing: limited or robust, with predefined steps and expected results, used where the risk calls for repeatability and traceability. - Evidence from the supplier: the developer's own testing and quality practices can be relied on where an assessment of the supplier supports it. Matching assurance effort to process risk Feature Process risk Fitting assurance activity Record Spell check in a document editor Not high Supplier evidence, no further testing Intended use and risk decision Dashboard that trends closed CAPAs Not high Unscripted: exploratory testing What was tested, by whom, result Electronic signature on batch release High Scripted: robust testing Steps, expected and actual results, conclusion Limit check that rejects out-of-range results High Scripted: robust testing, including failure cases Steps, expected and actual results, conclusion A reading of the final guidance's approach. The examples are ours; each team records its own risk decision. ## How does this relate to the QMSR? The Quality Management System Regulation, which amended 21 CFR Part 820 to incorporate ISO 13485:2016 by reference, took effect on 2 February 2026. Software used in the quality system still has to be validated for its intended use, with the effort proportionate to the risk. The CSA guidance describes how FDA expects that proportionate effort to be decided and recorded. ## What does CSA mean for AI-enabled quality tools? The same question applies: what does this feature do, and what happens if it is wrong? An AI feature that drafts text for a qualified person to review and approve sits at a different risk level from one that changes a record or a disposition by itself. Keeping the human decision in the workflow, and recording it, is what keeps the process risk low. AI adds one thing ordinary software does not: its behaviour can change when the model or the prompt changes. The assurance record should name the model version, and a change of model should go through change control with regression testing against a fixed set of cases. ## Common questions ### Does CSA replace computer system validation? CSA changes how validation effort is decided; it does not remove the requirement. Software used in production or the quality system must still be validated for its intended use. The guidance supersedes the section of FDA's 2002 General Principles of Software Validation that covered automated process equipment and quality system software. ### Does the CSA guidance apply to pharmaceutical manufacturers? It was issued for medical device production and quality system software. Drug manufacturers remain subject to 21 CFR Parts 210 and 211 and Part 11, and commonly apply the same risk-based approach through GAMP 5 Second Edition. ### Is unscripted testing acceptable to FDA? Yes, for features that are not high process risk. The guidance names ad hoc testing, error guessing and exploratory testing as acceptable assurance activities, with a record of what was tested, by whom, and the result. ## Sources - [Federal Register, 24 September 2025: Computer Software Assurance for Production and Quality System Software, final guidance, notice of availability](https://www.federalregister.gov/documents/2025/09/24/2025-18468/computer-software-assurance-for-production-and-quality-system-software-guidance-for-industry-and) - [Federal Register, 2 February 2024: Medical Devices; Quality System Regulation Amendments (QMSR final rule, effective 2 February 2026)](https://www.federalregister.gov/documents/2024/02/02/2024-01709/medical-devices-quality-system-regulation-amendments) - [FDA: Quality Management System Regulation (QMSR)](https://www.fda.gov/medical-devices/postmarket-requirements-devices/quality-management-system-regulation-qmsr) This article is general information for quality and regulatory professionals. It is not regulatory or legal advice. Check the source documents for their current status before acting on them. ## Topics - Computer software assurance - CSA - CSV - QMSR - GAMP 5 ## How QLogix helps - [QMS digitization](https://qlogix.ai/services/qms-digitization) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems) [All articles](https://qlogix.ai/blog) ## More from the blog - [Manufacturing science · 15 October 2026Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviationA result inside specification can still be the first point of a shift that ends in an out-of-specification batch three lots later. Control charts with alert limits at two sigma, action limits at three and Nelson's run rules see the shift while every result still passes. When the chart, the AI's reading of each signal and QA's disposition sit on one audited record, trending stops being an annual review exercise and becomes continued process verification.](https://qlogix.ai/blog/spc-signal-detection-gmp-trending) - [AI in quality · 15 October 2026Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by ageMost quality backlogs are worked in the order they arrived, which means the deviation on a sterile batch waits behind a labelling query from March. Triage by risk ranks every open item on patient and product impact, regulatory exposure, recurrence, due dates and age, says why in one line each, and puts QA's decision on each priority on record. The queue is then worked in the order that matters, and the order is defensible.](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization) - [AI in quality · 15 October 2026Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replacedA preventive interval copied from a manual is a guess. The work orders, downtime and calibration results a site already records contain the answer: mean time between failures, availability, whether failures are rising with age, and the interval that catches most of them before they happen. When the CMMS, the calibration module and the quality system share one record graph, the model reads itself, the engineer confirms, and the change goes through change control.](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp) ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # White papers: AI in the GxP quality system | QLogix Source: https://qlogix.ai/whitepapers # White papers One-page briefs for quality, regulatory and operations leaders on putting AI to work in a GxP quality system, each with its sources, charts and the controls that keep decisions with qualified people. Share a few details to download a paper. InvestigationsRoot cause analysisCAPA ## From deviation to defensible root cause AI-enabled investigations for GxP manufacturing Why inadequate investigations remain among FDA's most frequent observations, and how an AI-drafted analysis, hypothesis testing and an 11-point root cause check keep every decision with the investigator. Product: [QLogix CAPA management](https://qlogix.ai/products/deviation-capa-management) PDF · 1 page · October 2026 Download the PDF Agentic QMSArchitectureAI agents ## One quality system, one AI layer The case for an Agentic QMS What separate quality systems cost in validation, interfaces and oversight, and how one compliant record layer lets AI agents see the whole quality system and act only with a person's approval. Product: [QLogix Agentic QMS](https://qlogix.ai/products/agentic-qms) PDF · 1 page · October 2026 Download the PDF Quality risk managementFMEAICH Q9(R1) ## Risk assessments that stay true to the data Q Intelligence and the AI-enabled FMEA Quality risk management under ICH Q9(R1): occurrence and detection scored from the organization's own records, every Critical and High risk followed up, and the residual risk accepted on evidence. Product: [QLogix Risk assessment](https://qlogix.ai/products/quality-risk-management-fmea) PDF · 1 page · October 2026 Download the PDF Regulatory affairsNDA and BLAPost-approval changes ## From change control to filed dossier AI-enabled regulatory submissions on an Agentic QMS Post-approval changes screened for five markets, NDA and BLA dossiers checked against a 62-section eCTD template, and submission text in which every claim quotes its source. Product: [QLogix Regulatory submissions](https://qlogix.ai/products/regulatory-submissions) PDF · 1 page · October 2026 Download the PDF Complaint handlingInvestigationsBacklog ## From complaint to closed CAPA, with nothing left behind An Agentic QMS for complaints, investigations and backlog The reporting clocks a complaint can start, how AI triage hands each complaint to an investigation with its evidence, and how a risk-ranked backlog puts the oldest serious records first. Product: [QLogix Complaint handling](https://qlogix.ai/products/complaint-handling) PDF · 1 page · October 2026 Download the PDF LIMSCalibrationLogbooksCMMS ## Lab and plant, connected LIMS, calibration, logbooks and CMMS on one AI-enabled infrastructure How shared records turn GMP dependencies into rules: no result on an out-of-calibration instrument, calibration failures feeding maintenance reliability, and logbook reviews the system checks first. Product: [QLogix LIMS](https://qlogix.ai/products/lims) PDF · 1 page · October 2026 Download the PDF ValidationCSV and GAMP 5Project management ## Validation, digitized and on schedule Electronic validation with project management built in Eighteen templated validation documents on the V-model, electronic test execution with traceability, and a schedule that reports every day whether the validation project is on track. Product: [QLogix Validation](https://qlogix.ai/products/computer-system-validation) PDF · 1 page · October 2026 Download the PDF PM LibraryVisual inspectionFill-finish ## From particle to qualified inspection An AI-enabled particulate matter library for fill-finish Visual inspection is only as good as the defects people and machines are trained on. How a characterized, versioned library keeps inspector requalification and AVI challenge sets current with the particles the process actually produces. Product: [QLogix PM Library](https://qlogix.ai/products/pm-library) PDF · 1 page · October 2026 Download the PDF Each download link works once, for the paper requested. Your details go to the QLogix team only; see the [Privacy Policy](https://qlogix.ai/privacy). ## Talk it through with someone who has run a plant Bring your backlog, your systems map or one investigation that keeps coming back. We will tell you plainly what we would do and what we would leave alone. [Request a meeting](https://qlogix.ai/contact)[Email us](mailto:contact@qlogix.ai) In the first meeting we cover - Your current QMS and how it connects to MES, LIMS and ERP - Where deviations, CAPAs and change controls slow down - Whether a readiness assessment makes sense, and what it would cover --- # Request a meeting | QLogix Source: https://qlogix.ai/contact # Request a meeting Tell us about your site and what is on your mind. We will reply to set up a time to talk. Name Work email Company Role (optional) What would you like to look at? Website Send request --- # Site map | QLogix Source: https://qlogix.ai/sitemap # Site map Every public QLogix page. Each page is also available as plain Markdown at its address plus “.md”, for AI assistants and readers that prefer text. ## Company - [QLogix home](https://qlogix.ai/)[Markdown](https://qlogix.ai/index.md) - [Services](https://qlogix.ai/services)[Markdown](https://qlogix.ai/services.md) - [Products](https://qlogix.ai/products)[Markdown](https://qlogix.ai/products.md) - [Blog](https://qlogix.ai/blog)[Markdown](https://qlogix.ai/blog.md) - [White papers](https://qlogix.ai/whitepapers)[Markdown](https://qlogix.ai/whitepapers.md) - [Request a meeting](https://qlogix.ai/contact)[Markdown](https://qlogix.ai/contact.md) - [Site map](https://qlogix.ai/sitemap)[Markdown](https://qlogix.ai/sitemap.md) ## Services - [QMS digitization](https://qlogix.ai/services/qms-digitization)[Markdown](https://qlogix.ai/services/qms-digitization.md) - [AI-assisted investigations](https://qlogix.ai/services/ai-assisted-investigations)[Markdown](https://qlogix.ai/services/ai-assisted-investigations.md) - [Connected quality data](https://qlogix.ai/services/connected-quality-data)[Markdown](https://qlogix.ai/services/connected-quality-data.md) - [Validation of AI-enabled systems](https://qlogix.ai/services/validation-of-ai-enabled-systems)[Markdown](https://qlogix.ai/services/validation-of-ai-enabled-systems.md) - [Agentic workflow automation](https://qlogix.ai/services/agentic-workflow-automation)[Markdown](https://qlogix.ai/services/agentic-workflow-automation.md) ## Products - [Agentic QMS](https://qlogix.ai/products/agentic-qms)[Markdown](https://qlogix.ai/products/agentic-qms.md) - [Root cause investigation](https://qlogix.ai/products/root-cause-investigation)[Markdown](https://qlogix.ai/products/root-cause-investigation.md) - [CAPA management](https://qlogix.ai/products/deviation-capa-management)[Markdown](https://qlogix.ai/products/deviation-capa-management.md) - [Complaint handling](https://qlogix.ai/products/complaint-handling)[Markdown](https://qlogix.ai/products/complaint-handling.md) - [Batch record review](https://qlogix.ai/products/batch-record-review)[Markdown](https://qlogix.ai/products/batch-record-review.md) - [Change management](https://qlogix.ai/products/change-control)[Markdown](https://qlogix.ai/products/change-control.md) - [Document control](https://qlogix.ai/products/document-control)[Markdown](https://qlogix.ai/products/document-control.md) - [Risk assessment](https://qlogix.ai/products/quality-risk-management-fmea)[Markdown](https://qlogix.ai/products/quality-risk-management-fmea.md) - [Regulatory submissions](https://qlogix.ai/products/regulatory-submissions)[Markdown](https://qlogix.ai/products/regulatory-submissions.md) - [LIMS](https://qlogix.ai/products/lims)[Markdown](https://qlogix.ai/products/lims.md) - [Electronic lab notebook](https://qlogix.ai/products/electronic-lab-notebook)[Markdown](https://qlogix.ai/products/electronic-lab-notebook.md) - [MES (electronic batch records)](https://qlogix.ai/products/mes-electronic-batch-records)[Markdown](https://qlogix.ai/products/mes-electronic-batch-records.md) - [ERP](https://qlogix.ai/products/erp-materials-inventory)[Markdown](https://qlogix.ai/products/erp-materials-inventory.md) - [GMP logbooks](https://qlogix.ai/products/gmp-electronic-logbooks)[Markdown](https://qlogix.ai/products/gmp-electronic-logbooks.md) - [PM Library](https://qlogix.ai/products/pm-library)[Markdown](https://qlogix.ai/products/pm-library.md) - [Stability Study Management](https://qlogix.ai/products/stability-study-management)[Markdown](https://qlogix.ai/products/stability-study-management.md) - [Calibration management](https://qlogix.ai/products/calibration-management)[Markdown](https://qlogix.ai/products/calibration-management.md) - [CMMS](https://qlogix.ai/products/cmms)[Markdown](https://qlogix.ai/products/cmms.md) - [Validation](https://qlogix.ai/products/computer-system-validation)[Markdown](https://qlogix.ai/products/computer-system-validation.md) - [Project management](https://qlogix.ai/products/project-management)[Markdown](https://qlogix.ai/products/project-management.md) - [Operational excellence](https://qlogix.ai/products/operational-excellence)[Markdown](https://qlogix.ai/products/operational-excellence.md) - [Q Intelligence, Q Agent and agents](https://qlogix.ai/products/ai-for-pharma-quality)[Markdown](https://qlogix.ai/products/ai-for-pharma-quality.md) ## Blog - [Signal detection in GMP: how control charts with run rules catch a shift or a drift before it becomes a deviation](https://qlogix.ai/blog/spc-signal-detection-gmp-trending)[Markdown](https://qlogix.ai/blog/spc-signal-detection-gmp-trending.md) - [Backlog triage: ranking open deviations, CAPAs and change controls by risk instead of by age](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization)[Markdown](https://qlogix.ai/blog/backlog-triage-risk-based-prioritization.md) - [Predictive maintenance in GMP: how CMMS, calibration and reliability data decide when equipment is serviced, calibrated and replaced](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp)[Markdown](https://qlogix.ai/blog/predictive-maintenance-cmms-calibration-gmp.md) - [Connecting LIMS to investigations and CAPA: how AI speeds up OOS investigations and strengthens compliance](https://qlogix.ai/blog/lims-investigation-capa-integration-ai)[Markdown](https://qlogix.ai/blog/lims-investigation-capa-integration-ai.md) - [AI in deviation investigations: faster, more consistent root cause analysis](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis)[Markdown](https://qlogix.ai/blog/ai-deviation-investigation-root-cause-analysis.md) - [AI-assisted CAPA: from root cause to a verified effectiveness check](https://qlogix.ai/blog/ai-capa-management-effectiveness)[Markdown](https://qlogix.ai/blog/ai-capa-management-effectiveness.md) - [AI in document management and validation workflows for life sciences](https://qlogix.ai/blog/ai-document-management-validation-workflow)[Markdown](https://qlogix.ai/blog/ai-document-management-validation-workflow.md) - [FDA's seven-step credibility framework for AI, in plain terms](https://qlogix.ai/blog/fda-ai-credibility-framework)[Markdown](https://qlogix.ai/blog/fda-ai-credibility-framework.md) - [AI in drug manufacturing: where FDA stands and what to build for](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda)[Markdown](https://qlogix.ai/blog/ai-in-drug-manufacturing-fda.md) - [AI-enabled medical devices: what FDA's list and guidance show](https://qlogix.ai/blog/ai-enabled-medical-devices-fda)[Markdown](https://qlogix.ai/blog/ai-enabled-medical-devices-fda.md) - [Robotics in aseptic manufacturing: what Annex 1 expects](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1)[Markdown](https://qlogix.ai/blog/robotics-aseptic-manufacturing-annex-1.md) - [EU GMP Annex 22: what the draft AI annex asks of GMP manufacturers](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft)[Markdown](https://qlogix.ai/blog/eu-gmp-annex-22-ai-draft.md) - [Computer software assurance: what FDA's final guidance changes for validation](https://qlogix.ai/blog/computer-software-assurance-final-guidance)[Markdown](https://qlogix.ai/blog/computer-software-assurance-final-guidance.md) ## Legal - [Terms of Use](https://qlogix.ai/termsofuse)[Markdown](https://qlogix.ai/termsofuse.md) - [Privacy Policy](https://qlogix.ai/privacy)[Markdown](https://qlogix.ai/privacy.md) ## For AI assistants and search engines - [llms.txt](https://qlogix.ai/llms.txt) A short, link-rich summary of the site for AI assistants (llmstxt.org). - [llms-full.txt](https://qlogix.ai/llms-full.txt) The full text of every public page in one file. - [sitemap.xml](https://qlogix.ai/sitemap.xml) The XML site map for search engines. - [Blog RSS feed](https://qlogix.ai/blog/rss.xml) New and updated blog posts. --- # Terms of Use | QLogix Source: https://qlogix.ai/termsofuse # Terms of Use Last updated October 12, 2026 ## AGREEMENT TO OUR LEGAL TERMS We are QLogix ("Company," "we," "us," "our"), a company registered in California, United States at 5000 Birch Street, West Tower, Suite 3000, Newport Beach, CA 92660. We operate the website [https://qlogix.ai](https://qlogix.ai/) (the "Site"), as well as any other related products and services that refer or link to these legal terms (the "Legal Terms") (collectively, the "Services"). You can contact us by phone at (US)(714) 663-9012, email at [contact@qlogix.ai](mailto:contact@qlogix.ai), or by mail to 5000 Birch Street, West Tower, Suite 3000, Newport Beach, CA 92660, United States. These Legal Terms constitute a legally binding agreement made between you, whether personally or on behalf of an entity ("you"), and QLogix, concerning your access to and use of the Services. You agree that by accessing the Services, you have read, understood, and agreed to be bound by all of these Legal Terms. IF YOU DO NOT AGREE WITH ALL OF THESE LEGAL TERMS, THEN YOU ARE EXPRESSLY PROHIBITED FROM USING THE SERVICES AND YOU MUST DISCONTINUE USE IMMEDIATELY. We will provide you with prior notice of any scheduled changes to the Services you are using. The modified Legal Terms will become effective upon posting or notifying you by [contact@qlogix.ai](mailto:contact@qlogix.ai), as stated in the email message. By continuing to use the Services after the effective date of any changes, you agree to be bound by the modified terms. The Services are intended for users who are at least 18 years old. Persons under the age of 18 are not permitted to use or register for the Services. We recommend that you print a copy of these Legal Terms for your records. ## 1. OUR SERVICES The information provided when using the Services is not intended for distribution to or use by any person or entity in any jurisdiction or country where such distribution or use would be contrary to law or regulation or which would subject us to any registration requirement within such jurisdiction or country. Accordingly, those persons who choose to access the Services from other locations do so on their own initiative and are solely responsible for compliance with local laws, if and to the extent local laws are applicable. The Services are not tailored to comply with industry-specific regulations (Health Insurance Portability and Accountability Act (HIPAA), Federal Information Security Management Act (FISMA), etc.), so if your interactions would be subjected to such laws, you may not use the Services. You may not use the Services in a way that would violate the Gramm-Leach-Bliley Act (GLBA). ## 2. INTELLECTUAL PROPERTY RIGHTS ### Our intellectual property We are the owner or the licensee of all intellectual property rights in our Services, including all source code, databases, functionality, software, website designs, audio, video, text, photographs, and graphics in the Services (collectively, the "Content"), as well as the trademarks, service marks, and logos contained therein (the "Marks"). Our Content and Marks are protected by copyright and trademark laws (and various other intellectual property rights and unfair competition laws) and treaties in the United States and around the world. The Content and Marks are provided in or through the Services "AS IS" for your personal, non-commercial use or internal business purpose only. ### Your use of our Services Subject to your compliance with these Legal Terms, including the "PROHIBITED ACTIVITIES" section below, we grant you a non-exclusive, non-transferable, revocable license to: - access the Services; and - download or print a copy of any portion of the Content to which you have properly gained access, solely for your personal, non-commercial use or internal business purpose. Except as set out in this section or elsewhere in our Legal Terms, no part of the Services and no Content or Marks may be copied, reproduced, aggregated, republished, uploaded, posted, publicly displayed, encoded, translated, transmitted, distributed, sold, licensed, or otherwise exploited for any commercial purpose whatsoever, without our express prior written permission. If you wish to make any use of the Services, Content, or Marks other than as set out in this section or elsewhere in our Legal Terms, please address your request to: [contact@qlogix.ai](mailto:contact@qlogix.ai). If we ever grant you the permission to post, reproduce, or publicly display any part of our Services or Content, you must identify us as the owners or licensors of the Services, Content, or Marks and ensure that any copyright or proprietary notice appears or is visible on posting, reproducing, or displaying our Content. We reserve all rights not expressly granted to you in and to the Services, Content, and Marks. Any breach of these Intellectual Property Rights will constitute a material breach of our Legal Terms and your right to use our Services will terminate immediately. ### Your submissions Please review this section and the "PROHIBITED ACTIVITIES" section carefully prior to using our Services to understand the (a) rights you give us and (b) obligations you have when you post or upload any content through the Services. Submissions: By directly sending us any question, comment, suggestion, idea, feedback, or other information about the Services ("Submissions"), you agree to assign to us all intellectual property rights in such Submission. You agree that we shall own this Submission and be entitled to its unrestricted use and dissemination for any lawful purpose, commercial or otherwise, without acknowledgment or compensation to you. You are responsible for what you post or upload: By sending us Submissions through any part of the Services you: - confirm that you have read and agree with our "PROHIBITED ACTIVITIES" and will not post, send, publish, upload, or transmit through the Services any Submission that is illegal, harassing, hateful, harmful, defamatory, obscene, bullying, abusive, discriminatory, threatening to any person or group, sexually explicit, false, inaccurate, deceitful, or misleading; - to the extent permissible by applicable law, waive any and all moral rights to any such Submission; - warrant that any such Submission are original to you or that you have the necessary rights and licenses to submit such Submissions and that you have full authority to grant us the above-mentioned rights in relation to your Submissions; and - warrant and represent that your Submissions do not constitute confidential information. You are solely responsible for your Submissions and you expressly agree to reimburse us for any and all losses that we may suffer because of your breach of (a) this section, (b) any third party’s intellectual property rights, or (c) applicable law. ## 3. USER REPRESENTATIONS By using the Services, you represent and warrant that: (1) all registration information you submit will be true, accurate, current, and complete; (2) you will maintain the accuracy of such information and promptly update such registration information as necessary; (3) you have the legal capacity and you agree to comply with these Legal Terms; (4) you are not a minor in the jurisdiction in which you reside; (5) you will not access the Services through automated or non-human means, whether through a bot, script or otherwise; (6) you will not use the Services for any illegal or unauthorized purpose; and (7) your use of the Services will not violate any applicable law or regulation. If you provide any information that is untrue, inaccurate, not current, or incomplete, we have the right to suspend or terminate your account and refuse any and all current or future use of the Services (or any portion thereof). ## 4. USER REGISTRATION You may be required to register to use the Services. You agree to keep your password confidential and will be responsible for all use of your account and password. We reserve the right to remove, reclaim, or change a username you select if we determine, in our sole discretion, that such username is inappropriate, obscene, or otherwise objectionable. ## 5. PROHIBITED ACTIVITIES You may not access or use the Services for any purpose other than that for which we make the Services available. The Services may not be used in connection with any commercial endeavors except those that are specifically endorsed or approved by us. As a user of the Services, you agree not to: - Systematically retrieve data or other content from the Services to create or compile, directly or indirectly, a collection, compilation, database, or directory without written permission from us. - Trick, defraud, or mislead us and other users, especially in any attempt to learn sensitive account information such as user passwords. - Circumvent, disable, or otherwise interfere with security-related features of the Services, including features that prevent or restrict the use or copying of any Content or enforce limitations on the use of the Services and/or the Content contained therein. - Disparage, tarnish, or otherwise harm, in our opinion, us and/or the Services. - Use any information obtained from the Services in order to harass, abuse, or harm another person. - Make improper use of our support services or submit false reports of abuse or misconduct. - Use the Services in a manner inconsistent with any applicable laws or regulations. - Engage in unauthorized framing of or linking to the Services. - Upload or transmit (or attempt to upload or to transmit) viruses, Trojan horses, or other material, including excessive use of capital letters and spamming (continuous posting of repetitive text), that interferes with any party’s uninterrupted use and enjoyment of the Services or modifies, impairs, disrupts, alters, or interferes with the use, features, functions, operation, or maintenance of the Services. - Engage in any automated use of the system, such as using scripts to send comments or messages, or using any data mining, robots, or similar data gathering and extraction tools. - Delete the copyright or other proprietary rights notice from any Content. - Attempt to impersonate another user or person or use the username of another user. - Upload or transmit (or attempt to upload or to transmit) any material that acts as a passive or active information collection or transmission mechanism, including without limitation, clear graphics interchange formats ("gifs"), 1×1 pixels, web bugs, cookies, or other similar devices (sometimes referred to as "spyware"or"passive collection mechanisms"or"pcms"). - Interfere with, disrupt, or create an undue burden on the Services or the networks or services connected to the Services. - Harass, annoy, intimidate, or threaten any of our employees or agents engaged in providing any portion of the Services to you. - Attempt to bypass any measures of the Services designed to prevent or restrict access to the Services, or any portion of the Services. - Copy or adapt the Services' software, including but not limited to Flash, PHP, HTML, JavaScript, or other code. - Except as permitted by applicable law, decipher, decompile, disassemble, or reverse engineer any of the software comprising or in any way making up a part of the Services. - Except as may be the result of standard search engine or Internet browser usage, use, launch, develop, or distribute any automated system, including without limitation, any spider, robot, cheat utility, scraper, or offline reader that accesses the Services, or use or launch any unauthorized script or other software. - Make any unauthorized use of the Services, including collecting usernames and/or email addresses of users by electronic or other means for the purpose of sending unsolicited email, or creating user accounts by automated means or under false pretenses. - Use the Services as part of any effort to compete with us or otherwise use the Services and/or the Content for any revenue-generating endeavor or commercial enterprise. - Use the Services to advertise or offer to sell goods and services. - Sell or otherwise transfer your profile. ## 6. USER GENERATED CONTRIBUTIONS The Services does not offer users to submit or post content. ## 7. CONTRIBUTION LICENSE You and Services agree that we may access, store, process, and use any information and personal data that you provide following the terms of the Privacy Policy and your choices (including settings). By submitting suggestions or other feedback regarding the Services, you agree that we can use and share such feedback for any purpose without compensation to you. ## 8. THIRD-PARTY WEBSITES AND CONTENT The Services may contain (or you may be sent via the Site) links to other websites ("Third-Party Websites") as well as articles, photographs, text, graphics, pictures, designs, music, sound, video, information, applications, software, and other content or items belonging to or originating from third parties ("Third-Party Content"). Such Third-Party Websites and Third-Party Content are not investigated, monitored, or checked for accuracy, appropriateness, or completeness by us, and we are not responsible for any Third-Party Websites accessed through the Services or any Third-Party Content posted on, available through, or installed from the Services, including the content, accuracy, offensiveness, opinions, reliability, privacy practices, or other policies of or contained in the Third-Party Websites or the Third-Party Content. Inclusion of, linking to, or permitting the use or installation of any Third-Party Websites or any Third-Party Content does not imply approval or endorsement thereof by us. If you decide to leave the Services and access the Third-Party Websites or to use or install any Third-Party Content, you do so at your own risk, and you should be aware these Legal Terms no longer govern. You should review the applicable terms and policies, including privacy and data gathering practices, of any website to which you navigate from the Services or relating to any applications you use or install from the Services. Any purchases you make through Third-Party Websites will be through other websites and from other companies, and we take no responsibility whatsoever in relation to such purchases which are exclusively between you and the applicable third party. You agree and acknowledge that we do not endorse the products or services offered on Third-Party Websites and you shall hold us blameless from any harm caused by your purchase of such products or services. Additionally, you shall hold us blameless from any losses sustained by you or harm caused to you relating to or resulting in any way from any Third-Party Content or any contact with Third-Party Websites. ## 9. SERVICES MANAGEMENT We reserve the right, but not the obligation, to: (1) monitor the Services for violations of these Legal Terms; (2) take appropriate legal action against anyone who, in our sole discretion, violates the law or these Legal Terms, including without limitation, reporting such user to law enforcement authorities; (3) in our sole discretion and without limitation, refuse, restrict access to, limit the availability of, or disable (to the extent technologically feasible) any of your Contributions or any portion thereof; (4) in our sole discretion and without limitation, notice, or liability, to remove from the Services or otherwise disable all files and content that are excessive in size or are in any way burdensome to our systems; and (5) otherwise manage the Services in a manner designed to protect our rights and property and to facilitate the proper functioning of the Services. ## 10. PRIVACY POLICY We care about data privacy and security. Please review our Privacy Policy: [https://qlogix.ai/privacy](https://qlogix.ai/privacy). By using the Services, you agree to be bound by our Privacy Policy, which is incorporated into these Legal Terms. Please be advised the Services are hosted in the United States. If you access the Services from any other region of the world with laws or other requirements governing personal data collection, use, or disclosure that differ from applicable laws in the United States, then through your continued use of the Services, you are transferring your data to the United States, and you expressly consent to have your data transferred to and processed in the United States. ## 11. TERM AND TERMINATION These Legal Terms shall remain in full force and effect while you use the Services. WITHOUT LIMITING ANY OTHER PROVISION OF THESE LEGAL TERMS, WE RESERVE THE RIGHT TO, IN OUR SOLE DISCRETION AND WITHOUT NOTICE OR LIABILITY, DENY ACCESS TO AND USE OF THE SERVICES (INCLUDING BLOCKING CERTAIN IP ADDRESSES), TO ANY PERSON FOR ANY REASON OR FOR NO REASON, INCLUDING WITHOUT LIMITATION FOR BREACH OF ANY REPRESENTATION, WARRANTY, OR COVENANT CONTAINED IN THESE LEGAL TERMS OR OF ANY APPLICABLE LAW OR REGULATION. WE MAY TERMINATE YOUR USE OR PARTICIPATION IN THE SERVICES OR DELETE YOUR ACCOUNT AND ANY CONTENT OR INFORMATION THAT YOU POSTED AT ANY TIME, WITHOUT WARNING, IN OUR SOLE DISCRETION. If we terminate or suspend your account for any reason, you are prohibited from registering and creating a new account under your name, a fake or borrowed name, or the name of any third party, even if you may be acting on behalf of the third party. In addition to terminating or suspending your account, we reserve the right to take appropriate legal action, including without limitation pursuing civil, criminal, and injunctive redress. ## 12. MODIFICATIONS AND INTERRUPTIONS We reserve the right to change, modify, or remove the contents of the Services at any time or for any reason at our sole discretion without notice. However, we have no obligation to update any information on our Services. We will not be liable to you or any third party for any modification, price change, suspension, or discontinuance of the Services. We cannot guarantee the Services will be available at all times. We may experience hardware, software, or other problems or need to perform maintenance related to the Services, resulting in interruptions, delays, or errors. We reserve the right to change, revise, update, suspend, discontinue, or otherwise modify the Services at any time or for any reason without notice to you. You agree that we have no liability whatsoever for any loss, damage, or inconvenience caused by your inability to access or use the Services during any downtime or discontinuance of the Services. Nothing in these Legal Terms will be construed to obligate us to maintain and support the Services or to supply any corrections, updates, or releases in connection therewith. ## 13. GOVERNING LAW These Legal Terms and your use of the Services are governed by and construed in accordance with the laws of the State of California applicable to agreements made and to be entirely performed within the State of California, without regard to its conflict of law principles. ## 14. DISPUTE RESOLUTION ### Informal Negotiations To expedite resolution and control the cost of any dispute, controversy, or claim related to these Legal Terms (each a "Dispute"and collectively, the"Disputes") brought by either you or us (individually, a "Party"and collectively, the"Parties"), the Parties agree to first attempt to negotiate any Dispute (except those Disputes expressly provided below) informally for at least thirty (30) days before initiating arbitration. Such informal negotiations commence upon written notice from one Party to the other Party. ### Binding Arbitration If the Parties are unable to resolve a Dispute through informal negotiations, the Dispute (except those Disputes expressly excluded below) will be finally and exclusively resolved by binding arbitration. YOU UNDERSTAND THAT WITHOUT THIS PROVISION, YOU WOULD HAVE THE RIGHT TO SUE IN COURT AND HAVE A JURY TRIAL. The arbitration shall be commenced and conducted under the Commercial Arbitration Rules of the American Arbitration Association ("AAA") and, where appropriate, the AAA’s Supplementary Procedures for Consumer Related Disputes ("AAA Consumer Rules"), both of which are available at the [American Arbitration Association (AAA) website](http://www.adr.org). Your arbitration fees and your share of arbitrator compensation shall be governed by the AAA Consumer Rules and, where appropriate, limited by the AAA Consumer Rules. The arbitration may be conducted in person, through the submission of documents, by phone, or online. The arbitrator will make a decision in writing, but need not provide a statement of reasons unless requested by either Party. The arbitrator must follow applicable law, and any award may be challenged if the arbitrator fails to do so. Except where otherwise required by the applicable AAA rules or applicable law, the arbitration will take place in United States, California. Except as otherwise provided herein, the Parties may litigate in court to compel arbitration, stay proceedings pending arbitration, or to confirm, modify, vacate, or enter judgment on the award entered by the arbitrator. If for any reason, a Dispute proceeds in court rather than arbitration, the Dispute shall be commenced or prosecuted in the state and federal courts located in United States, California, and the Parties hereby consent to, and waive all defenses of lack of personal jurisdiction, and forum non conveniens with respect to venue and jurisdiction in such state and federal courts. Application of the United Nations Convention on Contracts for the International Sale of Goods and the Uniform Computer Information Transaction Act (UCITA) are excluded from these Legal Terms. In no event shall any Dispute brought by either Party related in any way to the Services be commenced more than one (1) years after the cause of action arose. If this provision is found to be illegal or unenforceable, then neither Party will elect to arbitrate any Dispute falling within that portion of this provision found to be illegal or unenforceable and such Dispute shall be decided by a court of competent jurisdiction within the courts listed for jurisdiction above, and the Parties agree to submit to the personal jurisdiction of that court. ### Restrictions The Parties agree that any arbitration shall be limited to the Dispute between the Parties individually. To the full extent permitted by law, (a) no arbitration shall be joined with any other proceeding; (b) there is no right or authority for any Dispute to be arbitrated on a class-action basis or to utilize class action procedures; and (c) there is no right or authority for any Dispute to be brought in a purported representative capacity on behalf of the general public or any other persons. ### Exceptions to Informal Negotiations and Arbitration The Parties agree that the following Disputes are not subject to the above provisions concerning informal negotiations binding arbitration: (a) any Disputes seeking to enforce or protect, or concerning the validity of, any of the intellectual property rights of a Party; (b) any Dispute related to, or arising from, allegations of theft, piracy, invasion of privacy, or unauthorized use; and (c) any claim for injunctive relief. If this provision is found to be illegal or unenforceable, then neither Party will elect to arbitrate any Dispute falling within that portion of this provision found to be illegal or unenforceable and such Dispute shall be decided by a court of competent jurisdiction within the courts listed for jurisdiction above, and the Parties agree to submit to the personal jurisdiction of that court. ## 15. CORRECTIONS There may be information on the Services that contains typographical errors, inaccuracies, or omissions, including descriptions, pricing, availability, and various other information. We reserve the right to correct any errors, inaccuracies, or omissions and to change or update the information on the Services at any time, without prior notice. ## 16. DISCLAIMER THE SERVICES ARE PROVIDED ON AN AS-IS AND AS-AVAILABLE BASIS. YOU AGREE THAT YOUR USE OF THE SERVICES WILL BE AT YOUR SOLE RISK. TO THE FULLEST EXTENT PERMITTED BY LAW, WE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, IN CONNECTION WITH THE SERVICES AND YOUR USE THEREOF, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT. WE MAKE NO WARRANTIES OR REPRESENTATIONS ABOUT THE ACCURACY OR COMPLETENESS OF THE SERVICES' CONTENT OR THE CONTENT OF ANY WEBSITES OR MOBILE APPLICATIONS LINKED TO THE SERVICES AND WE WILL ASSUME NO LIABILITY OR RESPONSIBILITY FOR ANY (1) ERRORS, MISTAKES, OR INACCURACIES OF CONTENT AND MATERIALS, (2) PERSONAL INJURY OR PROPERTY DAMAGE, OF ANY NATURE WHATSOEVER, RESULTING FROM YOUR ACCESS TO AND USE OF THE SERVICES, (3) ANY UNAUTHORIZED ACCESS TO OR USE OF OUR SECURE SERVERS AND/OR ANY AND ALL PERSONAL INFORMATION AND/OR FINANCIAL INFORMATION STORED THEREIN, (4) ANY INTERRUPTION OR CESSATION OF TRANSMISSION TO OR FROM THE SERVICES, (5) ANY BUGS, VIRUSES, TROJAN HORSES, OR THE LIKE WHICH MAY BE TRANSMITTED TO OR THROUGH THE SERVICES BY ANY THIRD PARTY, AND/OR (6) ANY ERRORS OR OMISSIONS IN ANY CONTENT AND MATERIALS OR FOR ANY LOSS OR DAMAGE OF ANY KIND INCURRED AS A RESULT OF THE USE OF ANY CONTENT POSTED, TRANSMITTED, OR OTHERWISE MADE AVAILABLE VIA THE SERVICES. WE DO NOT WARRANT, ENDORSE, GUARANTEE, OR ASSUME RESPONSIBILITY FOR ANY PRODUCT OR SERVICE ADVERTISED OR OFFERED BY A THIRD PARTY THROUGH THE SERVICES, ANY HYPERLINKED WEBSITE, OR ANY WEBSITE OR MOBILE APPLICATION FEATURED IN ANY BANNER OR OTHER ADVERTISING, AND WE WILL NOT BE A PARTY TO OR IN ANY WAY BE RESPONSIBLE FOR MONITORING ANY TRANSACTION BETWEEN YOU AND ANY THIRD-PARTY PROVIDERS OF PRODUCTS OR SERVICES. AS WITH THE PURCHASE OF A PRODUCT OR SERVICE THROUGH ANY MEDIUM OR IN ANY ENVIRONMENT, YOU SHOULD USE YOUR BEST JUDGMENT AND EXERCISE CAUTION WHERE APPROPRIATE. ## 17. LIMITATIONS OF LIABILITY IN NO EVENT WILL WE OR OUR DIRECTORS, EMPLOYEES, OR AGENTS BE LIABLE TO YOU OR ANY THIRD PARTY FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL, EXEMPLARY, INCIDENTAL, SPECIAL, OR PUNITIVE DAMAGES, INCLUDING LOST PROFIT, LOST REVENUE, LOSS OF DATA, OR OTHER DAMAGES ARISING FROM YOUR USE OF THE SERVICES, EVEN IF WE HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. NOTWITHSTANDING ANYTHING TO THE CONTRARY CONTAINED HEREIN, OUR LIABILITY TO YOU FOR ANY CAUSE WHATSOEVER AND REGARDLESS OF THE FORM OF THE ACTION, WILL AT ALL TIMES BE LIMITED TO THE LESSER OF THE AMOUNT PAID, IF ANY, BY YOU TO US OR $0.00 USD. CERTAIN US STATE LAWS AND INTERNATIONAL LAWS DO NOT ALLOW LIMITATIONS ON IMPLIED WARRANTIES OR THE EXCLUSION OR LIMITATION OF CERTAIN DAMAGES. IF THESE LAWS APPLY TO YOU, SOME OR ALL OF THE ABOVE DISCLAIMERS OR LIMITATIONS MAY NOT APPLY TO YOU, AND YOU MAY HAVE ADDITIONAL RIGHTS. ## 18. INDEMNIFICATION You agree to defend, indemnify, and hold us harmless, including our subsidiaries, affiliates, and all of our respective officers, agents, partners, and employees, from and against any loss, damage, liability, claim, or demand, including reasonable attorneys’ fees and expenses, made by any third party due to or arising out of: (1) use of the Services; (2) breach of these Legal Terms; (3) any breach of your representations and warranties set forth in these Legal Terms; (4) your violation of the rights of a third party, including but not limited to intellectual property rights; or (5) any overt harmful act toward any other user of the Services with whom you connected via the Services. Notwithstanding the foregoing, we reserve the right, at your expense, to assume the exclusive defense and control of any matter for which you are required to indemnify us, and you agree to cooperate, at your expense, with our defense of such claims. We will use reasonable efforts to notify you of any such claim, action, or proceeding which is subject to this indemnification upon becoming aware of it. ## 19. USER DATA We will maintain certain data that you transmit to the Services for the purpose of managing the performance of the Services, as well as data relating to your use of the Services. Although we perform regular routine backups of data, you are solely responsible for all data that you transmit or that relates to any activity you have undertaken using the Services. You agree that we shall have no liability to you for any loss or corruption of any such data, and you hereby waive any right of action against us arising from any such loss or corruption of such data. ## 20. ELECTRONIC COMMUNICATIONS, TRANSACTIONS, AND SIGNATURES Visiting the Services, sending us emails, and completing online forms constitute electronic communications. You consent to receive electronic communications, and you agree that all agreements, notices, disclosures, and other communications we provide to you electronically, via email and on the Services, satisfy any legal requirement that such communication be in writing. YOU HEREBY AGREE TO THE USE OF ELECTRONIC SIGNATURES, CONTRACTS, ORDERS, AND OTHER RECORDS, AND TO ELECTRONIC DELIVERY OF NOTICES, POLICIES, AND RECORDS OF TRANSACTIONS INITIATED OR COMPLETED BY US OR VIA THE SERVICES. You hereby waive any rights or requirements under any statutes, regulations, rules, ordinances, or other laws in any jurisdiction which require an original signature or delivery or retention of non-electronic records, or to payments or the granting of credits by any means other than electronic means. ## 21. CALIFORNIA USERS AND RESIDENTS If any complaint with us is not satisfactorily resolved, you can contact the Complaint Assistance Unit of the Division of Consumer Services of the California Department of Consumer Affairs in writing at 1625 North Market Blvd., Suite N 112, Sacramento, California 95834 or by telephone at (800) 952-5210 or (916) 445-1254. ## 22. MISCELLANEOUS These Legal Terms and any policies or operating rules posted by us on the Services or in respect to the Services constitute the entire agreement and understanding between you and us. Our failure to exercise or enforce any right or provision of these Legal Terms shall not operate as a waiver of such right or provision. These Legal Terms operate to the fullest extent permissible by law. We may assign any or all of our rights and obligations to others at any time. We shall not be responsible or liable for any loss, damage, delay, or failure to act caused by any cause beyond our reasonable control. If any provision or part of a provision of these Legal Terms is determined to be unlawful, void, or unenforceable, that provision or part of the provision is deemed severable from these Legal Terms and does not affect the validity and enforceability of any remaining provisions. There is no joint venture, partnership, employment or agency relationship created between you and us as a result of these Legal Terms or use of the Services. You agree that these Legal Terms will not be construed against us by virtue of having drafted them. You hereby waive any and all defenses you may have based on the electronic form of these Legal Terms and the lack of signing by the parties hereto to execute these Legal Terms. ## 23. CONTACT US In order to resolve a complaint regarding the Services or to receive further information regarding use of the Services, please contact us at: QLogix 5000 Birch Street West Tower, Suite 3000 Newport Beach, CA 92660 United States Phone: (US)(714) 663-9012 [contact@qlogix.ai](mailto:contact@qlogix.ai) EN SV TR PT-BR NB PL HU PT DE NO FI DA NL EL IT FR ES See also our [Privacy Policy](https://qlogix.ai/privacy). --- # Privacy Policy | QLogix Source: https://qlogix.ai/privacy # Privacy Policy Last updated October 12, 2026 This Privacy Notice for QLogix ("we," "us,"or"our"), describes how and why we might access, collect, store, use, and/or share ("process") your personal information when you use our services ("Services"), including when you: - Visit our website at [https://qlogix.ai](https://qlogix.ai/) or any website of ours that links to this Privacy Notice - Engage with us in other related ways, including any marketing or events Questions or concerns? Reading this Privacy Notice will help you understand your privacy rights and choices. We are responsible for making decisions about how your personal information is processed. If you do not agree with our policies and practices, please do not use our Services. If you still have any questions or concerns, please contact us at [contact@qlogix.ai](mailto:contact@qlogix.ai). ## SUMMARY OF KEY POINTS This summary provides key points from our Privacy Notice, but you can find out more details about any of these topics by clicking the link following each key point or by using our table of contents below to find the section you are looking for. What personal information do we process? When you visit, use, or navigate our Services, we may process personal information depending on how you interact with us and the Services, the choices you make, and the products and features you use. Learn more about personal information you disclose to us. Do we process any sensitive personal information? Some of the information may be considered "special"or"sensitive" in certain jurisdictions, for example your racial or ethnic origins, sexual orientation, and religious beliefs. We do not process sensitive personal information. Do we collect any information from third parties? We do not collect any information from third parties. How do we process your information? We process your information to provide, improve, and administer our Services, communicate with you, for security and fraud prevention, and to comply with law. We may also process your information for other purposes with your consent. We process your information only when we have a valid legal reason to do so. Learn more about how we process your information. In what situations and with which parties do we share personal information? We may share information in specific situations and with specific third parties. Learn more about when and with whom we share your personal information. How do we keep your information safe? We have adequate organizational and technical processes and procedures in place to protect your personal information. However, no electronic transmission over the internet or information storage technology can be guaranteed to be 100% secure, so we cannot promise or guarantee that hackers, cybercriminals, or other unauthorized third parties will not be able to defeat our security and improperly collect, access, steal, or modify your information. Learn more about how we keep your information safe. What are your rights? Depending on where you are located geographically, the applicable privacy law may mean you have certain rights regarding your personal information. Learn more about your privacy rights. How do you exercise your rights? The easiest way to exercise your rights is by submitting a [data subject access request](mailto:contact@qlogix.ai), or by contacting us. We will consider and act upon any request in accordance with applicable data protection laws. Want to learn more about what we do with any information we collect? Review the Privacy Notice in full. ## 1. WHAT INFORMATION DO WE COLLECT? ### Personal information you disclose to us In Short: We collect personal information that you provide to us. We collect personal information that you voluntarily provide to us when you register on the Services, express an interest in obtaining information about us or our products and Services, when you participate in activities on the Services, or otherwise when you contact us. Personal Information Provided by You. The personal information that we collect depends on the context of your interactions with us and the Services, the choices you make, and the products and features you use. The personal information we collect may include the following: - usernames - passwords - names - email addresses Sensitive Information. We do not process sensitive information. All personal information that you provide to us must be true, complete, and accurate, and you must notify us of any changes to such personal information. ## 2. HOW DO WE PROCESS YOUR INFORMATION? In Short: We process your information to provide, improve, and administer our Services, communicate with you, for security and fraud prevention, and to comply with law. We may also process your information for other purposes with your consent. We process your personal information for a variety of reasons, depending on how you interact with our Services, including: - To facilitate account creation and authentication and otherwise manage user accounts. We may process your information so you can create and log in to your account, as well as keep your account in working order. - To deliver and facilitate delivery of services to the user. We may process your information to provide you with the requested service. - To respond to user inquiries/offer support to users. We may process your information to respond to your inquiries and solve any potential issues you might have with the requested service. - To send administrative information to you. We may process your information to send you details about our products and services, changes to our terms and policies, and other similar information. - To request feedback. We may process your information when necessary to request feedback and to contact you about your use of our Services. - To send you marketing and promotional communications. We may process the personal information you send to us for our marketing purposes, if this is in accordance with your marketing preferences. You can opt out of our marketing emails at any time. For more information, see "WHAT ARE YOUR PRIVACY RIGHTS?" below. - To protect our Services. We may process your information as part of our efforts to keep our Services safe and secure, including fraud monitoring and prevention. - To identify usage trends. We may process information about how you use our Services to better understand how they are being used so we can improve them. ## 3. WHEN AND WITH WHOM DO WE SHARE YOUR PERSONAL INFORMATION? In Short: We may share information in specific situations described in this section and/or with the following third parties. We may need to share your personal information in the following situations: - Business Transfers. We may share or transfer your information in connection with, or during negotiations of, any merger, sale of company assets, financing, or acquisition of all or a portion of our business to another company. ## 4. DO WE USE COOKIES AND OTHER TRACKING TECHNOLOGIES? In Short: We may use cookies and other tracking technologies to collect and store your information. We may use cookies and similar tracking technologies (like web beacons and pixels) to gather information when you interact with our Services. Some online tracking technologies help us maintain the security of our Services and your account, prevent crashes, fix bugs, save your preferences, and assist with basic site functions. ### Google Analytics We may share your information with Google Analytics to track and analyze the use of the Services. To opt out of being tracked by Google Analytics across the Services, visit [https://tools.google.com/dlpage/gaoptout](https://tools.google.com/dlpage/gaoptout). For more information on the privacy practices of Google, please visit the [Google Privacy & Terms page](https://policies.google.com/privacy). ### Amplitude With your consent, we use Amplitude, a product analytics service, to understand how the Services, including the client portal, are used: pages viewed, clicks, forms started and submitted, and recordings of sessions. Values typed into forms are not recorded, the text of client records in the portal is masked before anything leaves your browser, and web addresses are sent without their query strings. Amplitude is not used unless you accept analytics cookies, and you can withdraw that consent at any time from Cookie settings at the foot of any page. For more information on the privacy practices of Amplitude, please visit the [Amplitude Privacy Notice](https://amplitude.com/privacy). ## 5. DO WE OFFER ARTIFICIAL INTELLIGENCE-BASED PRODUCTS? In Short: We offer products, features, or tools powered by artificial intelligence, machine learning, or similar technologies. As part of our Services, we offer products, features, or tools powered by artificial intelligence, machine learning, or similar technologies (collectively, "AI Products"). These tools are designed to enhance your experience and provide you with innovative solutions. The terms in this Privacy Notice govern your use of the AI Products within our Services. Use of AI Technologies We provide the AI Products through third-party service providers ("AI Service Providers"). As outlined in this Privacy Notice, your input, output, and personal information will be shared with and processed by these AI Service Providers to enable your use of our AI Products for purposes outlined in "WHEN AND WITH WHOM DO WE SHARE YOUR PERSONAL INFORMATION?" You must not use the AI Products in any way that violates the terms or policies of any AI Service Provider. Our AI Products Our AI Products are designed for the following functions: - AI search - AI bots - AI research - Text analysis - AI applications How We Process Your Data Using AI All personal information processed using our AI Products is handled in line with our Privacy Notice and our agreement with third parties. This ensures high security and safeguards your personal information throughout the process, giving you peace of mind about your data's safety. ## 6. HOW LONG DO WE KEEP YOUR INFORMATION? In Short: We keep your information for as long as necessary to fulfill the purposes outlined in this Privacy Notice unless otherwise required by law. We will only keep your personal information for as long as it is necessary for the purposes set out in this Privacy Notice, unless a longer retention period is required or permitted by law (such as tax, accounting, or other legal requirements). No purpose in this notice will require us keeping your personal information for longer than the period of time in which users have an account with us. When we have no ongoing legitimate business need to process your personal information, we will either delete or anonymize such information, or, if this is not possible (for example, because your personal information has been stored in backup archives), then we will securely store your personal information and isolate it from any further processing until deletion is possible. ## 7. HOW DO WE KEEP YOUR INFORMATION SAFE? In Short: We aim to protect your personal information through a system of organizational and technical security measures. We have implemented appropriate and reasonable technical and organizational security measures designed to protect the security of any personal information we process. However, despite our safeguards and efforts to secure your information, no electronic transmission over the Internet or information storage technology can be guaranteed to be 100% secure, so we cannot promise or guarantee that hackers, cybercriminals, or other unauthorized third parties will not be able to defeat our security and improperly collect, access, steal, or modify your information. Although we will do our best to protect your personal information, transmission of personal information to and from our Services is at your own risk. You should only access the Services within a secure environment. ## 8. DO WE COLLECT INFORMATION FROM MINORS? In Short: We do not knowingly collect data from or market to children under 18 years of age. We do not knowingly collect, solicit data from, or market to children under 18 years of age, nor do we knowingly sell such personal information. By using the Services, you represent that you are at least 18 or that you are the parent or guardian of such a minor and consent to such minor dependent’s use of the Services. If we learn that personal information from users less than 18 years of age has been collected, we will deactivate the account and take reasonable measures to promptly delete such data from our records. If you become aware of any data we may have collected from children under age 18, please contact us at [contact@qlogix.ai](mailto:contact@qlogix.ai). ## 9. WHAT ARE YOUR PRIVACY RIGHTS? In Short: You may review, change, or terminate your account at any time, depending on your country, province, or state of residence. Withdrawing your consent: If we are relying on your consent to process your personal information, which may be express and/or implied consent depending on the applicable law, you have the right to withdraw your consent at any time. You can withdraw your consent at any time by contacting us by using the contact details provided in the section "HOW CAN YOU CONTACT US ABOUT THIS NOTICE?" below. However, please note that this will not affect the lawfulness of the processing before its withdrawal nor, when applicable law allows, will it affect the processing of your personal information conducted in reliance on lawful processing grounds other than consent. Opting out of marketing and promotional communications: You can unsubscribe from our marketing and promotional communications at any time by clicking on the unsubscribe link in the emails that we send, or by contacting us using the details provided in the section "HOW CAN YOU CONTACT US ABOUT THIS NOTICE?" below. You will then be removed from the marketing lists. However, we may still communicate with you — for example, to send you service-related messages that are necessary for the administration and use of your account, to respond to service requests, or for other non-marketing purposes. ### Account Information If you would at any time like to review or change the information in your account or terminate your account, you can: - Log in to your account settings and update your user account. - Contact us using the contact information provided. Upon your request to terminate your account, we will deactivate or delete your account and information from our active databases. However, we may retain some information in our files to prevent fraud, troubleshoot problems, assist with any investigations, enforce our legal terms and/or comply with applicable legal requirements. Cookies and similar technologies: Most Web browsers are set to accept cookies by default. If you prefer, you can usually choose to set your browser to remove cookies and to reject cookies. If you choose to remove cookies or reject cookies, this could affect certain features or services of our Services. If you have questions or comments about your privacy rights, you may email us at [contact@qlogix.ai](mailto:contact@qlogix.ai). ## 10. CONTROLS FOR DO-NOT-TRACK FEATURES Most web browsers and some mobile operating systems and mobile applications include a Do-Not-Track ("DNT") feature or setting you can activate to signal your privacy preference not to have data about your online browsing activities monitored and collected. At this stage, no uniform technology standard for recognizing and implementing DNT signals has been finalized. As such, we do not currently respond to DNT browser signals or any other mechanism that automatically communicates your choice not to be tracked online. If a standard for online tracking is adopted that we must follow in the future, we will inform you about that practice in a revised version of this Privacy Notice. California law requires us to let you know how we respond to web browser DNT signals. Because there currently is not an industry or legal standard for recognizing or honoring DNT signals, we do not respond to them at this time. Global Privacy Control: We recognize and honor Global Privacy Control (GPC) signals. If you use a browser or extension that supports GPC, we will treat this as a valid request to opt out of the sale or sharing of your personal information for targeted advertising purposes under applicable state privacy laws, including the California Consumer Privacy Act (CCPA). When we detect a GPC signal from your browser, we will automatically apply your opt-out preference without requiring you to take any additional action. For more information about GPC and how to enable it, visit [globalprivacycontrol.org](http://globalprivacycontrol.org). ## 11. DO UNITED STATES RESIDENTS HAVE SPECIFIC PRIVACY RIGHTS? In Short: If you are a resident of California, Colorado, Connecticut, Delaware, Florida, Indiana, Iowa, Kentucky, Maryland, Minnesota, Montana, Nebraska, New Hampshire, New Jersey, Oregon, Rhode Island, Tennessee, Texas, Utah, or Virginia, you may have the right to request access to and receive details about the personal information we maintain about you and how we have processed it, correct inaccuracies, get a copy of, or delete your personal information. You may also have the right to withdraw your consent to our processing of your personal information. These rights may be limited in some circumstances by applicable law. More information is provided below. ### Categories of Personal Information We Collect The table below shows the categories of personal information we have collected in the past twelve (12) months. The table includes illustrative examples of each category and does not reflect the personal information we collect from you. For a comprehensive inventory of all personal information we process, please refer to the section "WHAT INFORMATION DO WE COLLECT?" Category Examples Collected A. Identifiers Contact details, such as real name, alias, postal address, telephone or mobile contact number, unique personal identifier, online identifier, Internet Protocol address, email address, and account name NO B. Personal information as defined in the California Customer Records statute Name, contact information, education, employment, employment history, and financial information NO C. Protected classification characteristics under state or federal law Gender, age, date of birth, race and ethnicity, national origin, marital status, and other demographic data NO D. Commercial information Transaction information, purchase history, financial details, and payment information NO E. Biometric information Fingerprints and voiceprints NO F. Internet or other similar network activity Browsing history, search history, online behavior, interest data, and interactions with our and other websites, applications, systems, and advertisements NO G. Geolocation data Device location YES H. Audio, electronic, sensory, or similar information Images and audio, video or call recordings created in connection with our business activities NO I. Professional or employment-related information Business contact details in order to provide you our Services at a business level or job title, work history, and professional qualifications if you apply for a job with us NO J. Education Information Student records and directory information NO K. Inferences drawn from collected personal information Inferences drawn from any of the collected personal information listed above to create a profile or summary about, for example, an individual’s preferences and characteristics NO L. Sensitive personal Information NO We may also collect other personal information outside of these categories through instances where you interact with us in person, online, or by phone or mail in the context of: - Receiving help through our customer support channels; - Participation in customer surveys or contests; and - Facilitation in the delivery of our Services and to respond to your inquiries. We will use and retain the collected personal information as needed to provide the Services or for: - Category G - As long as the user has an account with us ### Sources of Personal Information Learn more about the sources of personal information we collect in "WHAT INFORMATION DO WE COLLECT?" ### How We Use and Share Personal Information Learn more about how we use your personal information in the section, "HOW DO WE PROCESS YOUR INFORMATION?" Will your information be shared with anyone else? We may disclose your personal information with our service providers pursuant to a written contract between us and each service provider. Learn more about how we disclose personal information to in the section, "WHEN AND WITH WHOM DO WE SHARE YOUR PERSONAL INFORMATION?" We may use your personal information for our own business purposes, such as for undertaking internal research for technological development and demonstration. This is not considered to be "selling" of your personal information. We have not disclosed, sold, or shared any personal information to third parties for a business or commercial purpose in the preceding twelve (12) months. We will not sell or share personal information in the future belonging to website visitors, users, and other consumers. ### Your Rights You have rights under certain US state data protection laws. However, these rights are not absolute, and in certain cases, we may decline your request as permitted by law. These rights include: - Right to know whether or not we are processing your personal data - Right to access your personal data - Right to correct inaccuracies in your personal data - Right to request the deletion of your personal data - Right to obtain a copy of the personal data you previously shared with us - Right to non-discrimination for exercising your rights - Right to opt out of the processing of your personal data if it is used for targeted advertising (or sharing as defined under California’s privacy law), the sale of personal data, or profiling in furtherance of decisions that produce legal or similarly significant effects ("profiling") Depending upon the state where you live, you may also have the following rights: - Right to access the categories of personal data being processed (as permitted by applicable law, including the privacy law in Minnesota) - Right to obtain a list of the categories of third parties to which we have disclosed personal data (as permitted by applicable law, including the privacy law in California, Delaware, and Maryland) - Right to obtain a list of specific third parties to which we have disclosed personal data (as permitted by applicable law, including the privacy law in Minnesota and Oregon) - Right to obtain a list of third parties to which we have sold personal data (as permitted by applicable law, including the privacy law in Connecticut) - Right to review, understand, question, and depending on where you live, correct how personal data has been profiled (as permitted by applicable law, including the privacy law in Connecticut and Minnesota) - Right to limit use and disclosure of sensitive personal data (as permitted by applicable law, including the privacy law in California) - Right to opt out of the collection of sensitive data and personal data collected through the operation of a voice or facial recognition feature (as permitted by applicable law, including the privacy law in Florida) ### How to Exercise Your Rights To exercise these rights, you can contact us by submitting a [data subject access request](mailto:contact@qlogix.ai), by emailing us at [contact@qlogix.ai](mailto:contact@qlogix.ai), or by referring to the contact details at the bottom of this document. We will honor your opt-out preferences if you enact the [Global Privacy Control](https://globalprivacycontrol.org/) (GPC) opt-out signal on your browser. Under certain US state data protection laws, you can designate an authorized agent to make a request on your behalf. We may deny a request from an authorized agent that does not submit proof that they have been validly authorized to act on your behalf in accordance with applicable laws. ### Request Verification Upon receiving your request, we will need to verify your identity to determine you are the same person about whom we have the information in our system. We will only use personal information provided in your request to verify your identity or authority to make the request. However, if we cannot verify your identity from the information already maintained by us, we may request that you provide additional information for the purposes of verifying your identity and for security or fraud-prevention purposes. If you submit the request through an authorized agent, we may need to collect additional information to verify your identity before processing your request and the agent will need to provide a written and signed permission from you to submit such request on your behalf. ### Appeals Under certain US state data protection laws, if we decline to take action regarding your request, you may appeal our decision by emailing us at [contact@qlogix.ai](mailto:contact@qlogix.ai). We will inform you in writing of any action taken or not taken in response to the appeal, including a written explanation of the reasons for the decisions. If your appeal is denied, you may submit a complaint to your state attorney general. ### California "Shine The Light" Law California Civil Code Section 1798.83, also known as the "Shine The Light"law, permits our users who are California residents to request and obtain from us, once a year and free of charge, information about categories of personal information (if any) we disclosed to third parties for direct marketing purposes and the names and addresses of all third parties with which we shared personal information in the immediately preceding calendar year. If you are a California resident and would like to make such a request, please submit your request in writing to us by using the contact details provided in the section"HOW CAN YOU CONTACT US ABOUT THIS NOTICE?" ## 12. DO WE MAKE UPDATES TO THIS NOTICE? In Short: Yes, we will update this notice as necessary to stay compliant with relevant laws. We may update this Privacy Notice from time to time. The updated version will be indicated by an updated "Revised" date at the top of this Privacy Notice. If we make material changes to this Privacy Notice, we may notify you either by prominently posting a notice of such changes or by directly sending you a notification. We encourage you to review this Privacy Notice frequently to be informed of how we are protecting your information. ## 13. HOW CAN YOU CONTACT US ABOUT THIS NOTICE? If you have questions or comments about this notice, you may email us at [contact@qlogix.ai](mailto:contact@qlogix.ai) or contact us by post at: QLogix 5000 Birch Street West Tower, Suite 3000 Newport Beach, CA 92660 United States ## 14. HOW CAN YOU REVIEW, UPDATE, OR DELETE THE DATA WE COLLECT FROM YOU? Based on the applicable laws of your country or state of residence in the US, you may have the right to request access to the personal information we collect from you, details about how we have processed it, correct inaccuracies, or delete your personal information. You may also have the right to withdraw your consent to our processing of your personal information. These rights may be limited in some circumstances by applicable law. To request to review, update, or delete your personal information, please fill out and submit a [data subject access request](mailto:contact@qlogix.ai). EN SV TR PT-BR NB PL HU PT DE NO FI DA NL EL IT FR ES See also our [Terms of Use](https://qlogix.ai/termsofuse).