GxP news - QFINITY
QFINITY · News · Insights & Events

Shaping the regulated quality of tomorrow.

Articles and analysis from our own work across GxP, pharma, quality management, GAMP and AI - what is changing in the regulated environment, what it means in practice, and what is happening inside our company.

Latest articles

Events, case studies and insights.

Oliver Herrmann at the lectern of the main conference of the Biopharmaceutical Bioprocess Development Summit in Shanghai

One conference day in Shanghai, packed and inspiring in equal measure: a keynote and a panel at the Biopharmaceutical Bioprocess Development Summit, plus the deep-dive AI session at the AI for Pharma 2026 running in parallel. The question raised on the panel is one we currently meet in projects and committees alike: how do the three European drafts Chapter 4, Annex 11 and Annex 22 work together? This article shares the reading we gave on stage and looks into the background of shifting responsibilities, authority cases and criticality, as a working hypothesis based on the drafts and our observations.

Oliver Herrmann and Martin Heitmann under the entrance arch of AI for Pharma 2026 in Shanghai
TWO INVITATIONSInvited independently, on stage together: Oliver Herrmann and Martin Heitmann at the venue in Shanghai.

The occasion was unusual enough to be worth telling: a CMC and bioprocess summit put EU GMP regulation on its keynote program, and the AI for Pharma 2026 running in parallel booked the matching AI deep dive. Martin Heitmann and I had been invited independently of each other and only discovered it during preparation. Companies there want to know the criteria while there is still time to design for them. One day on site, tightly scheduled, and on the main conference panel the question that carries this article: how do all these data integrity requirements actually fit together?

The order behind the three drafts

Keynote slide: One Delivery, Three Rule Books - delivery, cargo, truck, autopilot and driver as an image for Chapter 4, Annex 11 and Annex 22
KEYNOTE SLIDEOne picture, three rule books: the cargo is the data (Chapter 4), the truck is the application and the road the qualified infrastructure (Annex 11), autopilot and guardrails belong to Annex 22, and a human stays at the wheel, carrying responsibility.

Our answer on the panel starts with where the legal act sits. Chapter 4 lives in the main part of the EU GMP Guide, and that is also where batch certification takes place: the Qualified Person certifies the batch on the basis of the records. That is not a system function but a legal act, and it is now meant to rest entirely on records, beyond documents in the traditional sense. This is why Chapter 4 addresses the data and its governance.

Annex 11 is the technical implementation. Its job is to give this legal act technical and functional trust: the computerized system in which the records are created, validated against its intended use and operated in a validated state. Annex 22, finally, inherits from both. It governs AI in critical GMP applications and presupposes the data and system control that Chapter 4 and Annex 11 have built. In the end, the three texts have to add up to a coherent control system, or one of them has not done its job.

One caveat belongs to every one of these statements: all three texts are drafts from the consultation. What will finally be published is open. Until then, this reading is a working hypothesis, based on our observations in the industry and on the EMA’s communication.

As early as 2022, the EMA concept paper on the Annex 11 revision announced that the FDA guidance on Computer Software Assurance, then available as a draft, would be examined: "This guidance and any implication will be considered with regards to aspects of potential regulatory relevance for GMP Annex 11." A guidance from the medical device world, written for production and quality system software and final in the meantime, is thus explicitly on the radar of the pharma revision. The convergence reaches across the Atlantic.

Exhibit 1: The person who moved

To test the ordering formula, follow the Qualified Person through the texts. The Annex 11 of 2011 named them explicitly: "…only Qualified Persons [to] certify the release of the batches". In the 2025 draft, they no longer appear. Instead, they now stand in the Chapter 4 draft: "All records should be available to the Qualified Person at the time of the release decision."

The move confirms the order: certification rests on the records, so the Qualified Person belongs in the chapter that governs the records. The system-side execution, meaning the signature in the system and the workflow behind it, remains a matter for Annex 11. On stage, we made the point in a single sentence:

The QP did not leave. The annex stopped being about them.

Exhibit 2: The numbers

The second test is quantitative. We measured the drafts against the 2011 versions, on the primary texts themselves:

1
Annex 11 grows from 5 to 19 pages. A statement of principles becomes a catalog of requirements that increasingly describes what control has to look like.
2
Chapter 4 grows from 9 to 17 pages and from 32 to 85 clauses. Both clause series are numbered without gaps; the chapter is still called "Documentation".
3
Four terms that appeared exactly zero times in 2011 now carry the chapter. Counted in the text: "governance" 19 times, "data integrity" more than twenty times, "lifecycle" 15 times, "criticality" 7 times.

The unassailable core of this measurement is the zero: a documentation chapter builds its foundation on terms it did not even know in 2011. The numbers show the direction, away from the document as a container, toward the data and its lifecycle as the object of control.

The test question both drafts ask: how strongly do these data influence the decision resting on them, and would you even notice an error?

Exhibit 3: Criticality has three readings and two axes

"Critical" appears several times across the drafts and does not mean the same thing three times over. The Chapter 4 draft defines data criticality in its glossary as "the degree of influence that data have on product safety as well as the regulatory compliance of processes, decisions and product quality". The same draft adds a second axis, detectability: would you see it if the data were wrong? The Annex 22 draft, in turn, calls applications critical when they have a "direct impact on patient safety, product quality or data integrity".

The third reading we took from a conference presentation: at the ISPE Pharma 4.0 conference in Barcelona in 2025, the EMA rapporteur explained the intention behind the critical-application concept along two axes, direct impact and detectability of the error. Detectability thus stands in both drafts, in Chapter 4 as in Annex 11 ("the likelihood of detection"), and additionally in the explained intention behind Annex 22; a single find becomes a pattern.

In practice, both converge on ALCOA++: criticality drives the rigor with which the ten attributes are demonstrated. And the second plus, Traceable, is the attribute-side counterpart of the detectability axis: it makes an error findable after the fact.

Exhibit 4: Existing requirements already apply to the use of AI

Anyone who considers the drafts a distant prospect should read two authority cases from this year. In April 2026, the FDA charged a manufacturer in a warning letter with "overreliance on artificial intelligence for your drug manufacturing operations". And in June 2026, the MHRA described AI-written inspection responses in its Inspectorate blog with "references to MHRA guidance that doesn’t exist", and drew the line that matters: "our concern isn’t whether you use AI; it’s whether your submissions are accurate, verifiable, and prepared under appropriate oversight".

Both authorities check the same thing: whether the evidence is accurate, verifiable and produced under appropriate oversight. The use of AI as such is not in question in either case. That is exactly the logic of the three drafts, applied before their finalization.

What follows from this

Oliver Herrmann at the microphone on the main conference panel in Shanghai
THE PANELThe question about the interplay of the requirements came from the main conference panel. This article’s answer is the worked-out version.

The consequence of our observations: the culture these drafts presuppose can be built today, with a data inventory, assigned criticality and a governance that puts the cross-cutting questions where they belong. Someone has to start, and what the first one builds sets the measure for every system that follows. That is exactly why these foundations belong at the overarching QA level: laid out once there, they carry across all systems instead of emerging by chance in whichever project comes first. QFINITY has supported this build-up since the ERES programs of the Part 11 era and from the core team of GAMP 5 Second Edition; we read the drafts before they become the rule.

And the sentence that drew the strongest reaction in Shanghai belongs at the end, because it explains why this control architecture exists in the first place:

Our industry is not machines serving patients. It is people serving people. The colleague who runs the bioreactor today may be the patient waiting for the vial tomorrow.

More on the foundation of this reading: our pages on the Annex 11 revision and the EU GMP Annex 22 track the state of the two drafts, and our analysis of PIC/S recommendation PI 006-4 shows how the same movement reaches the production level.

PI 006-4 Qualifizierung und Validierung - vom schrittweisen zum kontinuierlichen Nachweis - QFINITY

PIC/S has rewritten its recommendations on qualification and validation: PI 006-4 replaces the 2007 version on October 1, 2026. The real change sits beneath the chapters: validation is no longer a completed event but a continuously demonstrated state. That is precisely what gives modern ways of working a regulatory foundation, from a scientifically justified number of batches in process validation to ongoing verification in operation. Reading the new text with the vocabulary of 2007 can take you down the wrong path: scope and terminology have shifted. Computerized systems are explicitly out of scope; they belong to a different set of rules.

PI 006-4 "Recommendations on Qualification and Validation" is the Pharmaceutical Inspection Co-operation Scheme’s (PIC/S) recommendation on qualification and validation in pharmaceutical manufacturing; its participating authorities include the European inspectorates and the U.S. FDA. It covers the qualification of facilities, equipment and supporting utilities, process and cleaning validation, the validation of test methods, and special topics such as the verification of transportation and the validation of packaging for solid dose products. It enters into force on October 1, 2026, replaces the 2007 version and describes what inspectorates expect beyond Annex 15. It is not a legal instrument; as a guidance and training resource for GMP inspectors and the pharmaceutical industry, however, it shapes inspection practice directly.

Why is PI 006-4 a paradigm shift?

Four narrowly scoped topics from 2007 have become a lifecycle guide. The language, however, deserves particular attention: the words have stayed the same while the meaning underneath them has moved. The longer your validation practice reaches back, the more familiar the terms sound; they no longer carry their 2007 meaning.

1
Retrospective validation: gone. Once the lifeline for legacy processes. PI 006-4 describes it as "no longer considered an acceptable approach". Legacy processes need a gap analysis, a risk assessment and ongoing verification in line with today’s expectations.
2
Periodic revalidation of the process: no longer exists. Ongoing Process Verification "has taken the place of periodic revalidation". The state of control is demonstrated continuously; the fixed calendar cycle no longer applies. OPV is documented evidence of the state of control and therefore more than monitoring. What it replaces is periodic revalidation, not retrospective validation: two different removals.
3
Three batches: a transitional state. PIC/S itself marks the current Annex 15 wording as "transitory". The number of batches is scientifically justified and risk-based. The roadmap is public: the joint EMA and PIC/S Concept Paper from early 2026 announces a comprehensive review once the current targeted revision of Annex 15 is complete.
4
IQ, OQ and PQ: no longer the only way. Verification-based approaches built on good engineering practice, such as ASTM E2500, are explicitly named and can be applied where circumstances justify it. The classic qualification cascade remains the established route; it is simply no longer the only one.
5
Transportation is verified. The chapter is deliberately titled "Verification of Transportation". A transport route is movable and changeable; only stable equipment is qualified. PIC/S matches how the evidence is established to the object of evidence.

In addition, statistics moves from optional to expected: process capability indices, multivariate methods and predefined evaluation criteria belong in the protocols where risk warrants it, with subject matter experts and statisticians working side by side. Cleaning validation follows the same movement: the extent of the cleaning program is driven by a toxicological risk assessment, for example based on health-based exposure limits (HBEL); they make the actual hazard potential of an active substance the measure of how stringent cleaning needs to be.

Is periodic revalidation still written into your validation master plan?

Periodic revalidation gives way to Ongoing Process Verification, and its requirements are concrete. After the initial validation, once routine manufacturing begins, OPV runs until the process is discontinued. It monitors product quality and the critical parameters identified through risk assessment, and draws in quality system signals such as deviations and complaints; the trending should also be capable of detecting creeping change and special causes of variability. Results are reported regularly, normally at least once a year and, wherever possible, with objective statistical tools such as the process capability index Cpk, which measures how reliably the process stays within its specification limits; every report ends with a verdict: the process is in a state of control, or it is not. The effort scales with risk, and the reports may explicitly serve as a reference for the annual Product Quality Review. For most companies this means the data already exist. But what about the predefined criteria and the regular, documented verdict?

There is one question that lets you read PI 006-4 without stumbling: what exactly is demonstrated, and against what? Name the object and the reference point for every piece of evidence and you cannot take a wrong turn. Verifying the transport route then follows just as logically as the risk-based number of batches. Take only the familiar words with you, however, and verification turns into monitoring, the continuous demonstration of control turns into a calendar entry, and a conditional permission turns into a free choice. The same economy of thought runs through the document itself: evidence established once is referenced, not repeated.

  • FAT and SAT results may, with appropriate justification, reduce the scope of IQ and OQ.
  • PQ results feed into the risk assessment for process validation and can reduce its scope.
  • Ongoing Process Verification closes remaining minor gaps in the validation with justified additional testing.

What applies to computerized systems?

The two process worlds

Computerized systems are explicitly out of scope: their validation "is covered in the PIC/S GMP Guide Annex 11". That is a deliberate boundary between two process worlds. The Annex 15 process transforms material into product; it is tied to the plant, filed with the marketing authorization and changed through the variation procedure. The Annex 11 process originates in the business function and transforms information into records and decisions, from ERP processes to document control, QMS workflows or complaint handling. The tools of the production world do not transfer here: a deviation process has no batches and no process capability indices.

The difficulty starts where the two worlds overlap. An MES executes the filed manufacturing process; a LIMS manages testing against the approved methods. The system falls under Annex 11 while its content remains part of the marketing authorization world: the process world of Annex 15. Without that separation, you either test twice or leave a gap.

Three levels, one pattern

The boundary does not mean the systems world is spared the paradigm shift. It is ahead of it. Annex 11 requires computerized systems to be validated across the lifecycle and periodically evaluated in operation; the 2025 draft revision expands precisely that operational phase: reviews at risk-based intervals verify that the system remains in a validated state, and reveal when a system in motion drifts out of its defined parameters. The upcoming AI Annex 22 (2025 draft) takes it further: predefined metrics and acceptance criteria before testing, then regular performance monitoring of the model in operation and drift monitoring of the input data space. And the draft of the new Chapter 4 (Documentation, 2025) draws the same line at the data level: a data governance system should cover the entire data lifecycle, from creation through processing and archiving to destruction. PI 006-4 itself states that data governance should be considered in all aspects of qualification and validation.

From discrete steps to a continuous incline - stairs merging into a ramp
From steps to flowThe path stays the same; discrete steps become one continuous incline.

The movement is transatlantic, too, down to the authorship: the PIC/S working group behind the revision was most recently jointly chaired by Ireland’s HPRA and the U.S. FDA. The FDA anchored the lifecycle view in its Process Validation guidance back in 2011; Stage 3 is called Continued Process Verification there, and the 2015 revision of Annex 15 adopted that view. On the systems side, the FDA guidance on Computer Software Assurance shifts the effort from documentation to effective assurance; formally it covers the production and QMS software of medical device manufacturers, yet its approach draws attention well beyond that scope. Five documents in two years point in the same direction: the EMA and PIC/S Concept Paper on the Annex 15 revision, PI 006-4, the Annex 11 draft revision, the AI Annex 22 and the Chapter 4 draft. Production level, system level and data level converge on one pattern: quality is no longer proven at a point in time; it is demonstrated continuously.

And your computerized systems: under which set of rules do you validate?

QFINITY advises on both worlds and on the transition between them. Many of our clients work with the established methods today: classic validation campaigns, document-based evidence, the three-batch logic. We continue to advise on these paths without reservation; they remain acceptable under GMP, and which path fits is a question of maturity. Increasingly, and with the same conviction, we offer the continuous approaches: ongoing verification in operation, agile software development with a robust evidence trail and a risk-based number of batches. The first inquiries are already coming in. The deciding factor is where the organization stands: QFINITY covers past, current and emerging quality strategies and draws the line for every piece of evidence at its object. What is demonstrated, and against what? The answer determines whether the production rules apply (Annex 15; at the FDA, 21 CFR 211 with the process validation lifecycle) or the system rules (Annex 11; at the FDA, 21 CFR Part 11 and 211.68, and in the medical device world the CSA guidance), what depth of scrutiny is appropriate and what evidence your inspection requires. That evidence is what we deliver, and it is defensible.

PIC/S PI 006-4: Recommendations on Qualification and ValidationConcept Paper on the Revision of Annex 15 (EMA and PIC/S, 2026)

BPD Biopharmaceutical Bioprocess Development Summit - tour stops and dates across China, Singapore and the USA in 2026

We are glad to be at the 9th BPD Biopharmaceutical Bioprocess Development Summit in Shanghai!

The summit runs August 6-7, 2026 in Shanghai – China’s benchmark event for biopharmaceutical process development, from research to commercial manufacturing. Since 2017 it has connected the entire CMC value chain; the 2025 edition drew more than 5,500 attendees and over 160 speakers. The AI FOR PHARMA Conference 2026 runs alongside it.

QFINITY on stage

Oliver Herrmann (QFINITY) and Martin Heitmann (MH Consulting & Advisory) are pleased to speak at both conferences on August 6: at the BPD Summit itself on the shift in EU GMP regulation, and at the parallel AI FOR PHARMA Conference on how adaptive AI only becomes GMP-ready under control.

Shifting the EU GMP Regulatory Paradigm – Annex 11 and Annex 22 Drafts

Oliver Herrmann (QFINITY) and Martin Heitmann (MH Consulting & Advisory)

“Global Expansion & Globalization” track. For companies expanding into European markets, the regulatory ground is shifting: the revised EU GMP Annex 11 and the new Annex 22 draft reshape how computerized systems – and, increasingly, AI – are assessed in GxP environments. The talk maps out what quality, IT, and regulatory teams need to prepare for.

Adaptive AI Without Control is Not GMP – From Potential to Controlled Biopharma

Digital Compliance Architecture for Trustworthy Adaptive AI-Enabled Systems from Process Development to GxP Manufacturing

Oliver Herrmann (QFINITY) and Martin Heitmann (MH Consulting & Advisory)

AI FOR PHARMA Conference. Adaptive AI systems change their behavior as new data comes in – exactly what makes them powerful and hard to keep in check. Without robust control, that potential never reaches GMP. The talk outlines a digital compliance architecture that keeps adaptive, AI-enabled systems trustworthy and verifiable, from process development through to GxP manufacturing.

Meet us in Shanghai – we look forward to connecting in person.