# Ai Use-Case Selection Scorecard

Score each candidate from 1 to 5. Record evidence for every score.

| Dimension | Weight | Score | Evidence |
|---|---:|---:|---|
| User or business importance | 3 |  |  |
| Measurable value | 3 |  |  |
| Data or evidence readiness | 3 |  |  |
| Technical feasibility | 2 |  |  |
| User adoption readiness | 2 |  |  |
| Ability to verify outputs | 3 |  |  |
| Implementation effort | -2 |  |  |
| Privacy, fairness, security and safety risk | -3 |  |  |

## Decision rules

- Reject ideas with an unresolved legal, privacy or unacceptable-safety barrier.
- Do not use the total score alone; review the evidence and affected people.
- Prefer a narrow, measurable pilot over a broad transformation claim.
- Define a non-Ai baseline for comparison.

## Selected use case
- Candidate:
- Why it was selected:
- Pilot boundary:
- Success measure:
- Stop condition:
- Owner: