AI Compliance in the GMP Environment: Training with the Authors of the ISPE GAMP AI Guide

The integration of AI into GxP-regulated processes confronts quality systems with a fundamental question: how do you validate systems whose behaviour changes through learning? The classic requirements of EU GMP Annex 11 and 21 CFR Part 11 were developed for deterministic software. Adaptive, self-learning systems follow a different logic.

On 8 May 2026, Entourage hosted a four-hour live training that addresses this gap. Two of Europe's leading experts on AI compliance explained regulatory practice directly from their involvement in the relevant standards.

The Trainers

Oliver Herrmann, Chair of the ISPE GAMP European Steering Committee, has been an auditor for decades and a pioneer of IT compliance in the DACH region. Martin Heitmann, as co-lead of the ISPE GAMP Guide on Artificial Intelligence, is directly involved in the creation of the central industry standard.

Contents

The training covered four topic blocks:

Regulatory landscape. Where do Annex 11, the EU AI Act, and the upcoming EU GMP Annex 22 stand today? What exactly do inspectors check in AI systems, and which basis of interpretation applies when the frameworks contradict one another?

ISPE GAMP Guide AI. How does the new standard translate the abstract regulatory requirements into concrete implementation recommendations? What depth of validation does it require depending on the system category?

Validation strategies (CSV/CSA). How do you make the black-box nature of autonomous systems audit-proof? The training worked through the differences between Computer System Validation (CSV) and Computer Software Assurance (CSA) for AI applications directly using case examples.

Data Integrity & Governance. Lifecycle management for self-learning systems: what counts as a controlled change when model parameters shift through training? How are training data sets treated as a GMP document?

---

If you are interested in another run of this training or would like to tackle the topic of AI validation internally, get in touch with us.