Guide & tool · No registration
Define an AI use case your team can evaluate.
Use these nine prompts with the people who do the work, own the information and will support the result. This is a discussion aid, not a scored readiness assessment or production approval.
Add notes on a printed or saved copy. This page does not collect or store your answers.
- 01
The task & outcome
Who does the work today? What task should improve, and what would a useful result look like?
- 02
The baseline
What are the current time, cost, quality or service measures? What non-AI alternative should be compared?
- 03
Information & permission
Which documents, records or systems are needed? Who owns them and who may access each part?
- 04
Representative examples
Which typical, unusual and difficult cases should the solution handle? Keep a separate evaluation set.
- 05
Errors & consequences
What could go wrong? Which errors are tolerable, and which require a stop or escalation?
- 06
Human authority
What may the system suggest or do? Who approves consequential actions and resolves exceptions?
- 07
Integration & delivery
Which systems receive the output? What identity, interface, logging and recovery arrangements are needed?
- 08
Value & operating cost
How will you compare benefit with implementation, model, infrastructure, review and support costs?
- 09
Ownership & improvement
Who owns performance, monitoring, incidents, data updates and the decision to change or retire the solution?
Record the actions and owners.
Record what is known, what needs investigation, who owns each action and the evidence needed before proceeding. A useful outcome may be a focused proof, a simpler non-AI change, or a decision not to proceed.
Discuss a use case →