1. Describe your application
A short description is enough, you can refine later. The toolbox uses this to tailor the evaluation roadmap. The full flow takes roughly 20–30 minutes, and your progress is saved automatically in this browser.
2. Gateway screening
Three questions decide whether full evaluation is worth the effort. If benefit is low and risk is high, the framework recommends stopping here.
How the verdict is decided
- No clinical or patient need → STOP.
- Low benefit with high risk → STOP.
- Low benefit (any other risk level), or unclear need → PROCEED WITH CAUTION.
- Otherwise → PROCEED; the depth of evaluation scales with risk.
3. Customised evaluation roadmap
Ten dimensions, prioritised for your application. Each expands to evaluation items drawn from published evaluation frameworks and standards — FUTURE-AI, WHO guidance, the EU AI Act, NASSS and others (… items in all; click ⓘ on any item to see its source). Accept or reject each item to tailor your plan; rejected items are excluded. You'll mark each retained item's completion (Yes / Partial / No) in the next step.
4. Your evaluation plan
Work through the numbered checklist and mark each item's completion (Yes / Partial / No) as you assess your application. Click ⓘ on any item to see its ID, sub-aspect and source. Then save, share with your team, or hand to a reviewer.
Progress is saved automatically in this browser. To continue on another computer, or hand the checklist to a colleague, save the session file and load it there.