AI Use Case Prioritization
List your AI ideas, answer five questions for each, and see which to pilot first, which to fix before you start, and which to park. About 15 minutes, no sign-up, and nothing leaves your browser.
Five questions for each idea
Score each from 1 to 5. Risk counts the other way: 5 means a lot of it.
How much would it move a goal that matters: time saved, revenue, a better customer experience?
1 Nice to have5 Moves a key goalIs the data there, good enough, and yours to use?
1 We'd have to find or fix it5 Ready and approvedHow easy is it to build or buy?
1 A custom build5 Switch on a toolWill people use it, and does someone own it?
1 No owner, poor fit5 Owner and demandHow bad is it when the AI gets it wrong: privacy, compliance, customers, money?
1 Easy to catch, low harm5 Real harmScore your AI ideas
One idea per row, as specific as you can: "summarize support tickets for the weekly review" beats "AI for support". The map and plan below update as you go.
Where they land
Value against feasibility. Numbers match the plan.
Your plan
Best first. Priority is the average of value and feasibility, less a penalty for high risk.
Make it stick
- Score with the people who'd use it. Team readiness is a guess until they've weighed in.
- Pilot one thing at a time. One team, four to six weeks, one number to judge it by.
- Re-run it after each pilot. What you learn changes the scores for the rest.
Read Stop enabling individuals. Start enabling the organization. and Beyond Automation, or see where your team stands with the free PM + AI skill assessment (opens in a new tab).
Want help choosing and running your first AI pilot?