Insights
Short, sourced notes on AI and GMP quality systems. Each one says what is in force, what is a draft, and what is our own view.
An SOP changed. Which forms did you miss?
Why a revised SOP leaves forms, specifications and training records behind, and a short check to run before the effective date.
Draft Annex 22 in plain words: what it allows, what it excludes, and what to do now
For critical GMP uses the consultation draft describes only static, deterministic models and says generative AI should not be used there. Here is what that means for a QA team using AI on documents today.
Rules first, then AI: thirteen found in full, one in part, one missed
Same errors, same model. Fixed rules for the reference checks and the regulation given as text: 13 found in full, 1 in part, 1 missed, and no wholly false findings.
Twelve out of fifteen is not a pass: testing local AI on a demo quality system
Fifteen hidden errors, one local model, 3.4 minutes. What it found, what it missed and what it invented, with the full answer key.
The ten inspection questions, one by one
"Your data stays on site": what it proves, and what it does not
Can AI help with the Product Quality Review?
Looking for the source documents? See our learning resources.
New articles also appear on LinkedIn. Write to us if you want a topic covered.