Executive summary
A method for documenting what an AI business case assumes, testing the critical assumptions on real work, and freezing an honest register before appraisal.
What is inside
What the research pack covers
The asset walks a research lead through harvesting assumptions, grading them, ranking by materiality, designing a shadow pilot and research plan with kill thresholds, updating the register, challenging it, and freezing it. It ends with an assumption register, a source log, and a research question tracker, with a worked synthetic case.
- Turning a baseline and a readiness position into a governed set of AI business case assumptions
- Effect size, adoption, implementation cost, run cost, and cash conversion as ranges
- Confidence grades and dated sources for every number
- A shadow pilot that measures the tool on the organization's own work
- Kill thresholds that stop the case if the evidence is weak
- A challenge session and freeze before options are appraised