Where your executive team disagrees about AI without knowing it

AI Maturity Diagnostic, cohort mode

Each executive fills in the questionnaire alone, on a separate link, and none of them sees a report. What I look at is the spread between them.

In most teams I have seen, the CEO, the CFO and the head of operations describe three rather different companies when you ask them the same sixty questions, and none of them knows it, because nobody has ever put the answers side by side. That gap is usually the most expensive thing in the room. The questionnaire is only the instrument. What you are paying for is the ninety minutes in which we go through the differences that matter for a decision.

What a reading looks like

Here is the kind of thing that comes out. Everyone agrees that AI matters for the strategy. The CTO rates the data as ready; the two business owners rate it as not ready at all. Until one pilot that crosses departments has shown who owns the result and whether the data is actually there, a second wave of use cases would be money spent on an assumption.

What you receive

1. Your position on the scale, per dimension and overall, with the range within which it could reasonably move.
2. A written reading in ordinary business language, in Romanian or in English. I read every one of them before it goes out; nothing generated by a model reaches you unread.
3. Ninety minutes with the CEO, the executive team or the board, spent on the disagreements.
4. A short note with the next decisions, each with a name and a date.

The method, for those who want it

It is a structured questionnaire of about sixty items, twenty-five to thirty minutes, in Romanian or in English. Six dimensions are scored separately: technology, culture, strategy, execution, talent and governance. The result is a grade from 1 to 12, and the twelve grades follow the twelve phases of Călătoria AI. Each grade comes with a sensitivity range rather than a single number pretending to be precise. The scoring is deterministic; no model touches a score. There is also an internal consistency check, so that a flattering self-assessment shows up as a finding rather than as a result. Most maturity models have four or five stages, which is enough for a slide but not enough to tell you what to do next.

Level	What it is	When you use it
0	Deterministic scoring only. No model call, rules-based executive reading.	Data-sensitive environments, or a fast internal baseline.
1	Full run with AI-generated analysis.	The standard individual assessment.
2	Facilitated cohort: several executives on separate blind tokens, plus a debrief.	When the disagreement inside the team is the thing worth seeing.
3	Re-measurement at six months against the baseline cohort.	When you want to prove movement rather than claim it.

Data handling

Answers are stored on the server so results can be re-issued and compared over time. In blind mode the respondent sees no report and the analysis goes to the sponsor; respondents see an explicit consent notice before they start. A data handling note is supplied with every engagement, and I will send it before you commit if you want to read it first.

Ask for a cohort run

Three to six executives and a ninety-minute debrief. Individual runs are available if you ask.