Discovery

Methodology, not fine print

Every artifact linked here is built and evaluated on 100% synthetic, representative data by design. It is a capability and readiness reference, not a medical device.

Sprint 1 answers the four questions that decide whether to build at all - before a line of production code. It is the cheapest place to kill a bad idea and the fastest place to de-risk a good one.

The questions Discovery answers

• Is this the right opportunity? Score it, and name what was deprioritized and where AI is not the answer.
• Is the data ready? Grade provenance, licensing, privacy, coverage, and bias before building on it.
• What is the value? Model it on measured cost and an evidenced outcome, not an assumed one.
• What is the risk, and can it be bounded? Draw the safe envelope and confirm it can be enforced.

The Discovery evidence

Opportunity map Why this opportunity, what was deprioritized, and where AI is not the answer.

Data-readiness scorecard A graded readiness assessment, and the scorecard to run on your own data.

Value model The business case: measured run cost, illustrative outcome, and sensitivity.

Regulatory posture The compliance boundary and the medical-device line - the viability check.

Diagrams (C4) Context, container, and component views - the feasibility picture.

What you leave Discovery with

A go / no-go you can defend: the chosen opportunity, a graded data-readiness picture, a value model on your numbers, and a bounded risk envelope - or a documented reason not to build, which is just as valuable.

Next: PoC-to-Production.