Score every AI idea before anyone builds anything.
The AI Use Case Scorecard is a weighted ten-criteria rubric (max 60 points) for ranking AI use cases before you build anything. Rate each candidate workflow 1–5, count impact and frequency double, and the totals, not the loudest opinion in the room, pick your first initiative.
Ten rows. Five points each. No vibes.
| Criterion | Weight | What a 1 looks like | What a 5 looks like |
|---|---|---|---|
| Business impact | ×2 | Nobody would notice the improvement | Revenue, cost, or CX visibly moves |
| Workflow frequency | ×2 | Runs a few times a year | Runs daily, many times |
| Process clarity | ×1 | Different every time, undocumented | Written steps everyone follows |
| Data availability | ×1 | Lives in heads and inboxes | In systems AI can reach today |
| Implementation feasibility | ×1 | Needs new infrastructure first | First version shippable in weeks |
| Risk level | ×1 | Errors reach customers or money | Errors are internal and cheap to catch |
| Speed to validate | ×1 | Proof takes quarters | You would know within a week |
| Strategic value | ×1 | Dead end: helps nothing else | Builds data/integrations you reuse |
| Human review requirement | ×1 | Review takes as long as the work | Quick approval, then it ships |
| Integration complexity | ×1 | Touches five systems, all undocumented | One or two well-documented systems |
Copy this into a doc or spreadsheet, one column per candidate workflow. Or skip the setup: get the ready-to-use spreadsheet. Maximum weighted score is 60; anything above 45 deserves a serious look.
What the numbers are telling you
Common questions
What is a good score on the AI Use Case Scorecard? ›
The maximum weighted score is 60. Anything above 45 is a genuine Quick Win candidate: shortlist it. 30–44 means promising with a specific gap, usually process clarity or data availability. Fix the gap and rescore. Under 30 means it should not be your first AI initiative.
Why do business impact and workflow frequency count double? ›
Because value compounds with repetition. A high-impact workflow that runs daily creates more business value than an equally impressive one that runs quarterly. The double weighting keeps the totals honest about where AI actually pays off.
What if two candidates tie? ›
Break the tie on data availability: it is the gap that kills projects. The candidate whose data already lives in reachable systems ships weeks earlier and proves value faster than the one that needs a data cleanup first.
How many use cases should I score at once? ›
Five to ten real, specific workflows: "draft the weekly client report", not "marketing". Score them all in one sitting, rank the totals, and commit to the single top scorer for a quarter.
Get the scorecard as a ready-to-use spreadsheet. The weights are wired in, one column per workflow.
Free, no signup.