AI Use‑Case Prioritization Matrix & Scoring Tool

Purpose

This worksheet helps teams compare AI pilot candidates so you pick pilots with the best chance of safe clinical or operational impact. Use the scales below to score each candidate, apply the suggested weighting, and record a recommended next step. Save each completed form so your team can compare, sort, and discuss priorities.

Suggested scoring weights (example)

• Clinical impact: 30%
• Operational impact: 20%
• Feasibility: 20%
• Data readiness: 15%
• Regulatory & safety risk: -15% (penalty)

These weights are starting points — adjust them to match your organization’s Hungers (safety-first, equity, speed-to-value, etc.).

How to use

1. Fill one form per use case.
2. Score each criterion on the scales provided.
3. Compute a weighted score (manually or with your analytics tools) using your chosen weights. A high score indicates a higher-priority pilot.
4. Save and repeat. Then discuss the top candidates with governance and stakeholders before approval.

Sample scored use cases (illustrative)

Example A — Sepsis Early Warning: clinical impact 9, feasibility 6, data readiness 3, safety risk 7 → high priority but requires governance and prospective validation.

Example B — Appointment No-Show Prediction: clinical impact 3, operational impact 6, feasibility 8, data readiness 4, safety risk 2 → good quick-win pilot for operational ROI.

Governance checkpoints

• Clinical safety review for models affecting diagnosis/treatment.
• Privacy & security review for data access and linking.
• Regulatory review if model influences regulated decisions.
• Equity and bias assessment where outcomes vary by subgroup.
• Monitoring & rollback plan before any production rollout.

Next-step templates

After scoring, choose one of: proceed to a small pilot, run data readiness work, request governance review, or deprioritize. Use pilot templates to define success measures, monitoring, and a stop/rollback plan.