For mid-sized companies · 250–5 000 employees
Microsoft, Cisco, France Num and dozens of consultancies publish a readiness quiz. Ten minutes, twenty questions, a maturity level out of five. As a first conversation, they are fine. As the basis for a board decision, they fail for reasons that are structural — built into what a quiz is — and worth spelling out before you present one to a CFO.
Every answer is self-reported: how you believe your company uses AI, not what it does. The respondent who has read the most vendor decks scores highest, and the quiz has no way to check any of it against reality — no documents read, no job postings scanned, no public evidence consulted.
Two companies of the same size in different industries answer differently and get the same "next steps": train the team, define a governance framework, run a pilot. A score out of five carries no sector benchmark, no comparison with what the frontier in your industry is doing this quarter, and no opinion about what your sector has already stopped doing.
A CFO does not sign a maturity level. The quiz produces no ROI per initiative, no cost per step, no sequencing, no risk matrix — which is why the same companies retake the quiz every year and nothing moves between retakes.
The quiz exists to qualify you for the publisher’s pipeline: a score low enough to worry you routes you to their sales team. That does not make the score fake, but it explains the shape — reassuring enough to finish, vague enough to need a call. Nobody publishes a quiz whose honest output is "you do not need us".
Every quiz tells you what to add: a pilot, a policy, a committee. None tells you what to retire — the duplicated tooling, the abandoned pilot, the reporting nobody reads. Yet that stop list is what keeps the AI budget from growing faster than the results.
Retake the quiz in six months and the comparison is two scores. Nothing compounds: no baseline a roadmap can be tracked against, no record of what changed or why. Measurement that does not accumulate is a form, not a programme.
If nobody in the room has said a sentence about AI yet, take one. It costs nothing, it takes ten minutes, and it hands you shared vocabulary — the difference between “we should do something about AI” and a first argument about dimensions. The mistake is not taking the quiz; it is stopping there, with a score and no evidence, no benchmark and no money attached to anything.
The step past a quiz is a diagnostic that reads evidence instead of self-reports: public sources on your company and your sector, a benchmark against the frontier, an initiative portfolio with ROI, cost per step and a risk matrix — the parts a CFO can actually sign. That is what Become AI sells for 490 €, one-time, delivered in a day; the free preview on our site is the research half of the product, not a quiz. Like every business on NanoCorp, Become AI is built and run by AI agents, which is how the sector benchmarks stay current instead of aging into a slide deck.
The cost side of that comparison — what each option buys, including where a consultancy is the better answer — is on what an AI strategy costs.
Written 4 October 2026. The quizzes named are free self-assessments published by their vendors; the 490 € price is the one in our own checkout today. If a statement here stops being true, the page gets fixed the same day.