Pressure-test your AI project — before you build it.
A 9-part framework to align your team on what you're building, for whom, and why. Inspired by the Business Model Canvas, adapted for the realities of AI: data, integration, and stakeholders make or break delivery.
Force the hard conversations early. Surface gaps in data, skills, and value before engineering starts — when changing course is still cheap.
A one-page, exportable brief that a sponsor, engineer, or ops lead can read in three minutes and challenge meaningfully.
- 1Name your project
Set a clear, specific title in the sidebar — it anchors every AI suggestion.
- 2Frame the problem
Before the nine sections, define what you're actually solving, for whom, and the alternatives you'll have to beat.
- 3Draft each section
Write what you know. Leave gaps. The point is to surface unknowns, not to perform certainty.
- 4Turn claims into hypotheses
Under each section, write an if/then hypothesis with a kill metric and experiment. Update the status as you learn.
- 5Use AI & export
Suggest or Refine per section, Generate canvas to draft all nine, then export as Markdown with the hypothesis scorecard.