Self-assessment · 3 minutes
How ready is this AI decision?
Eight questions on value, readiness and authority. Your answers stay in your browser. A short result summary is saved in this tab; you choose whether to attach it to an enquiry. Nothing is sent automatically.
Your indication
Proceed
Move forward within a defined boundary.
Your answers describe a defined problem, workable foundations and clear authority. The next step is to confirm scope, controls and the conditions for release.
Investigate
Resolve the uncertainty before investing.
There is a case worth examining, but at least one material assumption is untested. Gather that evidence before committing to implementation.
Wait
A worthwhile idea may not be ready yet.
A prerequisite is missing: information, integration, operating ownership or capacity. Record it, assign an owner and set the condition for reconsidering.
Stop, for now
Do not build without a case.
The problem and its value are not yet defined well enough to justify investment. Clarify the decision first; a non-AI change may be the better answer.
By dimension
Resolve first
- Name one workflow, its owner and the outcome to change. “Use AI” is an ambition, not a decision boundary.
- Measure the current process first, so any benefit can be demonstrated rather than assumed.
- Define what success is worth and who is accountable for it before comparing solutions.
- Confirm sources, quality and permission. Assess personal or confidential information with your privacy adviser before any trial.
- Map the systems of record, integration points and who will operate, monitor and support the system.
- Confirm an accountable sponsor, budget and team time before committing to delivery.
- Assign a named person with the time, information and authority to review, override and stop the system.
- Match evaluation and approval to the consequence: high-impact decisions need stronger evidence and human review before action.
No single answer stands out as a blocker. Confirm the scope, controls and conditions for release.
An illustrative self-assessment based on your answers. It is not advice, a certification or an assessment of your organisation. If you request a review, you can choose whether to include this summary. Nothing is sent automatically.
Beyond a self-check
What a Review adds.
A self-assessment reflects one person’s view. The Review tests it with the people who own the outcome, examines the evidence and ends with a recommendation leadership can explain.
A recommendation. With its reasoning attached.
- The decision
- Should this workflow use AI?
- The boundary
- Defined process. Named owner. Human oversight.
- The evidence
- Value, readiness, risk and operating constraints.
Do not build a solution without a case.
The evidence does not justify further investment in this proposed use of AI, or the risks exceed the agreed boundary.
Next actionDocument the reasons and consider non-AI alternatives.
A worthwhile opportunity may not be ready.
A prerequisite is missing: data access, operating ownership, capacity or a dependency the organisation must resolve.
Next actionRecord the dependency, owner and condition that triggers reconsideration.
Resolve the uncertainty before investing.
A material assumption remains untested. Gather the missing evidence before making an implementation commitment.
Next actionDefine the evaluation question, representative cases and acceptance criteria.
Move forward within a defined boundary.
The agreed evidence supports a bounded next step, with named ownership and explicit operating controls.
Next actionConfirm the scope, approval authority and conditions for release.
Illustrative format. Not a client result, template promise or approved recommendation.