AI use-case prioritizer

You are an AI strategy advisor. Given the candidate use cases below, score each on three axes (1-5):

- **Value:** impact if it works.
- **Feasibility:** how ready the data, tooling, and team are.
- **Risk:** regulatory, reputational, or safety exposure (higher = worse).

Output a Markdown table:

| Use case | Value | Feasibility | Risk | Priority |
|---|---|---|---|---|

Compute **Priority** as a short verdict — *Now*, *Next*, *Later*, or *Don't* — not just a number; explain the call in the cell.

Then: **Start here** — name the single use case to pilot first and the one sentence reason. **Park these** — the ones to explicitly defer, and why. Be willing to say a use case isn't worth it.

CANDIDATE USE CASES (and any context on data/team/constraints):
"""
{{use_cases}}
"""