An AI readiness score is a point-in-time snapshot, not a grade, forecast or benchmark, and its single number hides which of the ten dimensions is dragging it down.
A board asks "what's our AI readiness score" and expects a single tidy answer, the way they would ask for a credit rating or an exam result. It is neither of those things, and treating it as though it were is where most of the confusion about these scores starts.
Before anyone puts the number in a board pack, it is worth being precise about what it actually represents, because the number survives long after the context that produced it is forgotten.
A readiness score is a snapshot: aggregated answers across a set of measurable dimensions of current practice, taken at one point in time, combining what staff report about their own habits with whatever artefacts can be checked against those reports. It describes where an organisation currently stands. It does not evaluate whether that is good, in any moral sense, only where the dimension sits relative to a defined scale.
It is not a maturity certification with anything to display. It is not meaningfully comparable to another organisation's score unless both were assessed with the same methodology at the same rigour, which is rarer than the comparison implies. It is not a pass or fail threshold: nothing changes automatically at 50 or 70. And it is not a forecast. A high score today says nothing certain about capability in six months, because the underlying practice can move in either direction faster than most people expect.
The honest answer is communicability, not precision. A single figure is the only version of this that fits on one page of a board report and can be tracked over time without a methodology lecture attached. The cost is that the average hides which dimension is dragging it down. An organisation scoring 60 because everything is mediocre looks identical, on the headline number, to one scoring 60 because two dimensions are strong and two are close to zero. Those are different organisations with different first moves, and the single number cannot tell them apart.
A number out of 100 is a conversation starter, not a verdict.
The trade-off is worth naming plainly rather than apologising for. Compression is what makes the artefact usable at board level at all: no board has the appetite for ten separate dimension scores with no summary, and no Executive Director wants to explain a methodology in the five minutes allotted to the item. The price of that usability is that the figure is less precise than it looks, and the responsible move is to say so every time it is presented, not just the first time.
Treat the top-line score as the invitation to read the ten-dimension breakdown behind it, never as a substitute for reading it.
No external statistic cited; this article defines methodology and scope rather than presenting evidence, so nothing in the source register was directly applicable.
If this is the question on your desk, a thirty-minute call tells you whether the service fits, or that you do not need us yet.