Fifteen scenarios, scored on judgment rather than on how much you use AI. You will see where you stand, what your ceiling is, and the one dimension that moves you fastest.
No account. No email to see your score.
Closing your weakest dimension is worth more than improving all five a little. The result page names which one and by how much.
The assessment costs nothing and asks for nothing. No account, no card, no email to see your result. You get your score whether or not you ever speak to us.
The companies building AI models cannot credibly grade how much you rely on their products. Measurement has to come from somewhere that does not sell you the tool.
Five fixed dimensions, scored one to four against a published rubric. Same questions, same scale, same scoring for everyone. How an answer becomes a score is written down in full; the answer key itself goes to reviewers on request rather than onto this page, because a key you have read is a score you chose.
McKinsey now calls AI fluency the next foundation of US economic competitiveness, and defines it as knowing when and how to use AI, and then how to verify and improve the results.
McKinsey, August 2026.
The same fifteen scenarios, aggregated by department. You see which teams are strong, where the risk concentrates, and what to do about it.
Verification carries the lowest average across the organization, at 2.0 of 4. That is the column to act on.
Illustrative figures, not a real organization.
Five dimensions, weighted equally. This is the whole rubric.
Knowing where AI helps, where it fails, and when not to use it.
Directing AI clearly with context, constraints, and intent.
Checking AI output for accuracy before relying on it.
Using AI within policy, privacy, and risk boundaries.
Building AI into real workflows so gains compound.
No trick questions and no obviously-right answer. Every option is something a competent person might genuinely choose.
You're on a deadline. An AI tool gives you a specific market figure you need for a slide and cites a named, real-looking source. You have 20 minutes left.
What actually demonstrates good judgment here?
Each answer maps to a level. The highest-scoring answer is often not the most cautious or the most thorough one.