• Publication Date: July 30, 2026
• Effective Date: October 1, 2026

The latest revision of PIC/S PI 006 represent a significant evolution in the approach to process validation (PV). Rather than relying on a predefined number of validation batches as the primary basis for demonstrating process robustness, the revised guidance promotes a science- and risk-based lifecycle approach.

The revised framework emphisizes process understanding, quality risk management, ongoing monitoring, and continued process verification thoughout the product lifecycle, reflecting the principles established in ICH Q9.

Why does this matter?

The revised approach provides manufacturers with greater flexibility in designing and justifying their process validation strategies. However, this flexibility comes with a greater expectation that the validation approach is scientifically justified, appropriately risk-based, and supported by documented process knowledge.

For companies preparing for validation activities or reassessing existing validation strategies, this may be an opportunity to:

• Re-evaluate whether a predefined three-batch approach remains appropriate for each process.
• Strengthen the connection between process understanding, QRM and validation strategy.
• Ensure that development and manufacturing data provide an adequate scientific basis for validation decisions.
• Define an appropriate continued process verification (CPV) strategy rather than viewing validation as a one-time exercise.
• Reassess previous validations in light of the lifecycle approach.

RJW Perspective

The move away from the traditional three-batch paradigm should not be interpreted as “fewer batches are now acceptable.” Rather, the revised approach places greater responsibility on manufacturers to demonstrate why their validation strategy is appropriate for the specific process, product and level of process understanding.

This may be particularly relevant for companies introducing new products, transferring processes to/from a CDMO, or reassessing established manufacturing processes. A science- and risk-based approach can provide greater flexibility, but it also requires a robust rationale, appropriate supporting data, and a clear connection between development knowledge, quality risk management and ongoing process monitoring.

• FDA Warning Letter issuance date: April 2026
• Company: Purolea Cosmetic Lab

The FDA inspection of Purolea was not an audit of an AI system. It was an audit of the company’s ability to know what it was doing. However, the investigator’s findings brought the use of AI into the spotlight.

Look at the sequence described in the Warning Letter:

• The company used AI to create specifications, procedures and master production records.
• The company apparently relied on AI to determine which FDA requirements needed to be applied.
• When FDA told the firm that process validation was required, the response was essentially: the AI had never told them that.
• At the same time, FDA identified fundamental failures involving the Quality Unit, testing, procedures, batch-record review and process validation.

Apart from the individual observations, the Warning Letter raises a very interesting question: What happens when an organization outsources its regulatory judgment to a machine?

• AI can: retrieve → summarize → draft → compare → suggest
• But the regulated organization must still: understand → challenge → decide → approve → own

And this distinction may be one of the most important lessons from the Warning Letter.

This is different from simply saying “AI needs human oversight”.

It is about organizational competence and accountability in an AI-enabled environment.

AI can be an incredibly powerful tool when used properly, it can speed up work, facilitate access to enormous amounts of information, help identify patterns, support document development and make many activities more efficient.

There is no reason for regulated organizations not to benefit from these capabilities. The issue is not AI, but how AI is used.

Consider a simple question: “What does FDA require for process validation?”.

If an organization asks AI that question and uses the answer as a starting point for further assessment, challenge and discussion, AI is doing exactly what a good tool should do: supporting the organization’s capabilities.

But if the organization accepts the answer without someone who understands process validation independently assessing it, the problem is no longer simply that AI might be wrong. It becomes a quality-system issue.

The most concerning aspect of the inspection may therefore be what the response to FDA revealed: not simply a gap in regulatory knowledge, but the inability of the organization to critically assess the information generated by AI.

And that represents a very different, and potentially much higher, risk.

Undoubtedly, companies will use more and more AI. And they should. AI can accelerate the creation of documents and provide access to an enormous amount of regulatory information. It can help professionals work faster. It can help them work smarter.

But AI cannot assume responsibility for a GMP decision.  This is where the discussion moves beyond AI governance and into the fundamentals of a mature Quality organization.

Companies should ask themselves whether their use of AI is strengthening the capabilities of their Quality organization or replacing the knowledge needed to challenge the output.

The question is therefore not simply: Is AI being used under human oversight?

The more important question is: Does the organization remain capable of exercising informed human judgment?

That is ultimately the real AI challenge for regulated companies.

• Publication date: July 28, 2026
• FDA – Generic Peptide Products – CMC & Regulatory

What happened?

On July 28, FDA withdrew its 2021 guidance, ANDAs for Certain Highly Purified Synthetic Peptide Drug Products That Refer to Listed Drugs of rDNA Origin, stating that it no longer reflects the Agency’s current scientific thinking.

At the same time, FDA issued 17 revised draft Product-Specific Guidances (PSGs) for generic peptide products.

FDA explains that the revised PSGs reflect the Agency’s regulatory experience, new scientific information and global regulatory developments. The Agency also indicates that it plans to revise the withdrawn guidance.

What caught our attention?

The withdrawal of the 2021 guidance is important, but we think the 17 PSGs themselves may be the more interesting part of this update.

When looking across the PSGs, we noticed that FDA is providing more specific information on some of the CMC questions that can be particularly challenging for peptide products, including:

• impurity reporting and identification thresholds;
• how impurities in the test product should be compared with those in the RLD;
• how the results of those comparisons may be used when establishing specifications; and
• recommendations that appear to be more consistent across different dosage forms of the same peptide.

For example, the revised PSG for glucagon recommends a NMT 1.0% limit for new impurities, while the general guidance has referenced a NMT 0.5% threshold. This is a good example of why applying a single general threshold to peptide products may not always be appropriate. The product-specific recommendation may provide a different and more targeted approach.

We also noticed that some of the recommendations appear to be moving closer to approaches already familiar to developers in Europe. We would be cautious about calling this an adoption of the European approach, but the similarities are worth watching.

Why does it matter?

For companies developing peptide ANDAs, these changes may have implications beyond the individual products covered by the 17 PSGs.

The revised PSGs provide a useful indication of how FDA is thinking about peptide characterization and control today, and may therefore be relevant when considering analytical strategies, impurity specifications and the overall CMC development approach for other peptide products.

For ongoing projects, this raises a practical question:

Should the current development strategy be revisited in light of what FDA is now recommending?

That may be particularly relevant where development decisions were made using the 2021 guidance as the primary reference.

RJW Perspective

We see this as more than a change in FDA documentation.

The combination of withdrawing the 2021 guidance and issuing 17 detailed PSGs suggests that FDA is continuing to refine its approach to generic peptides as scientific knowledge and regulatory experience evolve.

The level of detail in the new PSGs is particularly interesting. Rather than simply providing broad recommendations, FDA is giving developers more insight into how it expects certain peptide-specific CMC issues to be evaluated.

The next step will be to see how these recommendations are applied in practice and how the replacement for the 2021 guidance ultimately frames the broader approach.

For companies with peptide programs, the question may therefore not simply be “What changed?” but “What can these PSGs tell us about where FDA is heading?”

What should developers watch?

1. The replacement for the 2021 guidance
   FDA’s 2026 Guidance Agenda indicates that the Agency plans to revise the withdrawn guidance this year. The July 28 announcement, however, does not provide a specific publication date.
2. How the new impurity recommendations are applied
   Particularly where product-specific recommendations differ from the general thresholds previously used as a reference.
3. The relationship between the test product and the RLD
   The revised PSGs provide additional insight into how FDA expects impurity profiles and other product characteristics to be evaluated.
4. Consistency across dosage forms
   The recommendations for different dosage forms of the same peptide may provide useful insight into how FDA is approaching the products as a broader family rather than as completely separate development programs.
5. Convergence with other regulatory approaches
   Some of the revised recommendations appear increasingly consistent with approaches seen in other regions. It will be interesting to see whether this develops into a more harmonized approach over time.

Bottom line: The guidance may have been withdrawn, but the 17 PSGs give us a useful window into FDA’s current thinking on generic peptide development.

For companies with peptide programs, they may be worth a closer look—not only for the products specifically covered, but also for what they may tell us about the direction FDA is taking.