Where it started
Management at a medtech company wanted AI in device development within the year, and the budget was there. The department of 80 engineers had fourteen ideas, two earlier pilots that never made it into operation, and a QA function that had stopped both.
The strategic question
Which initiatives should run, in what order, and which should be left alone?
What we found
The scoping mapped where the engineers’ time actually goes, and which AI tools were already in use, approved or not. Each idea was tested against what works in operation today, not in a demo, and mapped to where it touches the QMS, documentation and validation.
# | Initiative | Hours freed / yr | Time to operation | Touches the QMS | Status |
|---|---|---|---|---|---|
1 | Drafting and review of design documents, human sign-off | ~2,000 | 2–3 months | No | Started |
2 | Traceability check across requirements, risk file and test protocols | ~5,000 | 6–9 months | Yes, tool validation | Prepared with QA |
3 | Test protocol drafts from requirements | ~1,500 | 4–6 months | Partly | Queued |
4 | Supplier document review | ~2,800 | 3–4 months | No | Queued |
5 | Impact analysis of design changes across the design history file | ~8,000 | 18–24 months | Yes, validated system | Full ambition |
— | Nine further ideas | — | — | — | Parked, each with a reason |