An AI readiness assessment scores an organization. The good ones look at data quality and access, infrastructure and integration, strategy and use cases, governance and risk, and the skills on hand. Cisco’s AI Readiness Index and Microsoft’s assessment are typical of the category, and consultancies run their own. If your question is whether the organization can stand a deployment up, that’s the right instrument, and nothing here replaces it.
The question the category can’t answer is why results stay uneven after the deployment is up. Usage climbs, the infrastructure is fine, the training happened, and the outcomes still depend on who’s using the tool and how. McKinsey’s State of AI survey (November 2025) found 88% of organizations using AI in at least one function and 39% seeing any effect on earnings. The readiness scores of the 88% didn’t predict membership in the 39%.
Where the missing variable lives
It lives during use. In the largest analysis of human-AI collaboration to date (Vaccaro, Almaatouq & Malone, Nature Human Behaviour; 106 experiments), the human-AI combination underperformed the better of the two alone on average. The split was by task type: on decision tasks the combination lost ground, and on creation tasks it gained when the person brought domain expertise. Same tools, same organization, different outcome, depending on what the person brought to the exchange. A readiness score can’t see that, because it isn’t a property of the organization. It’s a state in the person, and it moves.
The H(x)AI model names that state x, the mental model that couples a person’s capability with the tool’s, and describes four conditions that keep it high: bringing your judgment in because it’s the scarce input; engaging with the output instead of accepting it; holding your professional identity steady; and keeping your picture of what the tool can do calibrated. Those four are what “readiness” means at AIRAA. They’re states, which is why an assessment taken once tells you little and a practice repeated tells you a lot.
How to look at it
For yourself, one question before the next substantive AI request: what am I withholding from this that I already know? If the honest answer is “most of what makes me valuable,” the readiness score of your organization is not your problem.
For a team, watch outputs rather than attitudes. Could this piece of work have come from anyone with the same subscription? The AIRAA app teaches the four conditions in nine short modules and, as its next release, adds a four-question noticing instrument that reports which condition went quiet in a single episode. For a leadership team that wants to see the pattern across a group, the workshop does that in a half day.
Two instruments, two questions. Can the organization run AI? Ask the readiness assessment. Will the people get value from it once it’s running? That’s the during-use question, and it’s the one AIRAA is built for.
