Company environment · Synthetic API · FDA inspection
A drug-ingredient plant after an FDA warning
Commercial ingredient facing an FDA warning letter. Start at the top to see what the company does, then the jobs AI models were given here and how each scored.
1What does this company do?
A chemical plant that makes active ingredients other companies turn into medicines.
Its ingredient is already sold worldwide. A harmful impurity was found in it, and the FDA issued a warning letter.
2Where is the product on its way to patients?
The ingredient is on the market and under recall pressure.
- 1Discovery
- 2Lab & animal studies
- 3Human trials
- 4Filing & review
- 5Approved
- 6On the market
- The product is here
The plant must explain how the impurity formed, prove its test method, reconcile safety reports and complaints, justify recalls and audit its suppliers.
3What is happening right now?
It is 15 January 2019 and the warning-letter response is due.
Every answer must face the FDA’s citations directly, without inventing missing lab data.
4Which jobs were the models given?
6 jobs, each graded against a specialist’s checklist.
Scores are checks passed out of checks in the list. Each model ran each job once.
5–7Evidence, work and verdict
This company’s files stay private.
It is built from a real product’s records, so its files, runs and checklists aren’t published. To follow a job all the way down to each check, open one of the public environments.
Havenor Therapeutics (fictional, fully published) →
Broad Institute (real, public FDA record) →