What 'AI-ready' actually means — and how to measure it

Ask ten leaders what ‘AI-ready’ means and you will get ten answers. One means familiarity with the latest tools. Another means having taken a course. A third means a vague confidence that their people ‘get it’. That ambiguity is not a small problem. You cannot build, fund, hire for or prove something you cannot define — and right now, the most important quality in the labour market is almost entirely undefined.

The definition problem

Most definitions of AI-readiness fail in one of two directions. They are either too shallow — knowing what a model is, or being able to name three tools — or too abstract to measure, like ‘an AI mindset’. Shallow definitions are easy to game and mean little to an employer. Abstract ones cannot be compared between two people, which makes them useless for decisions.

A useful definition has to survive contact with a hiring manager. It has to let one person be compared honestly to another, and it has to be hard to fake. That rules out most of what passes for AI-readiness today.

A test for readiness

Here is the test we propose. Readiness is not whether someone has watched a course or can recite a definition. It is whether they can demonstrate, under observation, that they can use AI to reach a real outcome — and explain their reasoning. Demonstrated, not declared.

This reframes readiness from a noun you possess to a verb you perform. It is closer to how we assess a driver or a surgeon than how we assess a certificate. And it has a useful property: a demonstration can be scored, and a score can be compared.

Readiness you can only describe is a slogan. Readiness you can demonstrate and score is a standard.

Why measurement changes everything

The moment readiness becomes measurable, three things change at once. The individual gets an honest signal of where they actually stand, rather than the false comfort of a completed course. The institution gets a comparable number it can move — student by student, team by team — instead of a mission statement. And the employer gets something they can trust enough to act on.

Measurement also exposes the access gap. When you can see readiness clearly, you can see exactly who is being left out — which regions, which languages, which backgrounds — and you can direct help there first instead of averaging the problem away.

The obvious objections

• ‘You can't measure something this fuzzy.’ You can measure performance on real tasks; that is what every practical exam already does.
• ‘A score reduces a person to a number.’ A score is a floor, not a ceiling — a credible starting point that gets a capable person in the room.
• ‘People will just train to the test.’ If the test is a genuine demonstration of a useful skill, training to it is simply learning the skill.

What a standard unlocks

A shared, measurable standard for AI-readiness is the missing layer underneath the whole AI economy. It lets a student prove themselves to an employer they have never met. It lets an institution show the value it adds. It lets a government see where the divide is widening and act before it hardens. None of that is possible while ‘AI-ready’ remains a feeling rather than a number.

That is the work: not another course, but a credible standard the world can agree to read.