Most organisations that tell me they are AI ready are describing their software. They have licences, a policy document, a training day in the calendar and a few enthusiastic early adopters. None of that is readiness. It is activity. And I have watched activity like that fall apart the moment a governor asks a hard question, a parent complains, or a new model arrives and changes what the tools can do.
So here is the definition I work to, and the test I use to check it.
What being AI ready actually means
Being AI ready means your organisation can make, communicate and hold good decisions about AI, consistently, without depending on any one person. That is the whole definition. It says nothing about which tools you own, how many people use them, or whether your data is in the cloud. Those things matter, but they are inputs. Readiness is the capacity to decide well, and to keep deciding well as the technology moves.
Notice what the definition rules out. An organisation with one brilliant AI lead who holds everything in their head is not ready, because when they leave, so does the readiness. An organisation with a beautifully written policy that staff have never read is not ready, because a decision nobody knows about is not a decision, it is a document. And an organisation where most of the workforce quietly use AI every day, while leadership has not agreed what it is for, is not ready either. It is exposed.
I define readiness as a capacity rather than a state because AI does not stand still long enough for a state to mean anything. If readiness were a checklist, you would complete it and be unready again within a term. A capacity keeps working when the ground moves.
Why the question matters now
The question matters now because usage has run far ahead of leadership, and the gap between the two is where the risk lives. AI has already arrived in your organisation, on people's phones, in their inboxes and in tools nobody approved. The live question is whether leadership has caught up with what its own people are already doing.
The National Education Union's State of Education survey, published in April 2026 from responses by more than nine thousand teachers in English state schools, found that 76 per cent now use AI tools for their day-to-day work, up from 53 per cent a year earlier. In the same survey, 49 per cent said their school had no AI policy for staff or students at all. Read those two figures together. Three in four teachers are using AI, and half of them are doing it in an organisation that has not yet decided what it thinks.
That pattern is not unique to schools. Cisco's AI Readiness Index, published in October 2025 from a survey of eight thousand senior leaders across thirty markets, found that only around 13 per cent of organisations qualified as what it calls Pacesetters, the group that consistently turns AI activity into value. Almost all of those Pacesetters, 99 per cent, had a defined AI roadmap. Among everyone else, the figure was 58 per cent. The difference between the two groups was not enthusiasm for AI. It was whether the organisation had decided where it was going.
So the question "are we AI ready?" is really the question "has our leadership caught up with our own workforce?" For most organisations, the honest answer is not yet.
The Readiness Test: three questions
The Readiness Test is three questions an organisation must be able to answer consistently, from the top to the front line, before it can call itself AI ready: What is AI for here? What may I do with it? Who decides when it goes wrong? I use these questions because they are quick to ask and hard to fake. A leadership team can usually answer all three in the boardroom. The test is whether a teacher in a Year 8 classroom, or a finance officer, or a newly appointed middle leader, would give the same answers.
What is AI for here?
The first question tests whether your organisation has a purpose for AI, rather than a collection of tools. When I ask a leadership team this question, the strongest answer I hear is a single sentence that connects AI to something the organisation already cares about: giving teachers time back for the parts of teaching only humans can do, or removing the administrative drag that keeps senior staff away from the work they were appointed for. The weakest answer is a list of products. If the purpose cannot be said in a sentence, it cannot be shared, and if it cannot be shared, every department will invent its own.
What may I do with it?
The second question tests whether your decisions have reached the people who act on them. Most organisations I work with have a policy. Far fewer have staff who could tell you, without looking it up, what they are allowed to put into an AI tool, what they are not, and what they must check before they use the output. Readiness here is not the existence of rules. It is the confidence of the person using AI at four o'clock on a Thursday, with nobody to ask, who knows the answer anyway.
Who decides when it goes wrong?
The third question tests accountability, and it is the one leadership teams most often fail. AI will produce something wrong, biased or inappropriate at some point in every organisation that uses it. That is not a reason to avoid it, any more than the certainty of a safeguarding incident is a reason to close a school. But it does mean someone has to own the response. When I put this question to a room, the tell is a pause followed by several people looking at each other. If the answer is not immediate, the organisation is not ready, however good its tools.
What the evidence says about readiness
The research is consistent on one point: readiness is a leadership variable, not a technology variable. McKinsey's State of AI survey, published in August 2026 from 1,719 respondents across 97 countries, found that nearly nine in ten organisations now use AI regularly in at least one business function, yet only 6 per cent qualified as high performers, meaning they could attribute significant value to it. The organisations that did were twice as likely as the rest to report that senior leaders demonstrated visible commitment to AI initiatives, and nearly three quarters of them had fundamentally redesigned how work gets done rather than bolting AI onto existing processes.
Cisco's index tells the same story from a different direction. Its Pacesetters were four times more likely to move a pilot into production, and far more likely to have thought about the risks: 87 per cent said they were highly aware of AI-specific threats, against 42 per cent of everyone else. Awareness of the downside is a marker of readiness, not a sign of hesitation.
I read this evidence as confirmation of something I see every week. The organisations that get value from AI are not the ones that adopted it first or bought the most. They are the ones whose leaders made decisions, said them out loud, and changed the way work happened as a result.
What I See in Practice
Across the leadership teams I work with, the organisations that turn out to be least ready are often the ones that look most advanced from the outside. They have the highest usage figures and the most visible enthusiasm, and that is precisely the problem. Usage without decisions is not maturity. It is a set of habits forming in the dark, and habits are much harder to reshape than a blank page.
The pattern I see most often runs like this. A few early adopters start using AI and get real gains. Word spreads. Leadership notices, feels behind, and responds with the two things that are easiest to buy: a licence and a training day. Usage jumps. Then the questions start arriving that no licence can answer. A parent asks whether their child's work is being marked by a machine. A member of staff pastes something they should not have into a tool nobody approved. A department produces something with AI that another department would never have signed off. At that point the organisation discovers that it has adopted AI without ever having decided anything about it.
I recognise this pattern because I have been on both sides of it. As an assistant headteacher, and later as Director of Digital Strategy in further education, I led technology change in organisations that were expected to move fast, and I saw at first hand how easy it is to confuse rollout with readiness. What worked was not more technology. It was fewer, clearer decisions, made by the leadership team, communicated until people were tired of hearing them, and revisited on a schedule.
That is also what the strongest organisations I work with now have in common. They are not the most sophisticated users. They are the ones that can answer the three questions above from any seat in the building.
The mistakes that keep organisations unready
The mistakes that keep organisations unready are almost never technical; they are decisions that were delegated, deferred or never made. Four come up again and again.
The first is treating a policy as a strategy. A policy tells people what they may not do. It does not tell them what the organisation is trying to achieve, so the most cautious reading wins and the ambitious staff quietly go around it.
The second is outsourcing readiness to the most enthusiastic person in the building. Every organisation has one. They are valuable, and they are also a single point of failure. If your readiness lives in one person's head, you do not have readiness. You have a risk with a name.
The third is measuring usage and calling it maturity. Dashboards showing how many staff logged in this month are seductive, and they measure the wrong thing. High usage in an organisation that has not decided its purpose is a warning sign, not a success metric.
The fourth is waiting for certainty. Leaders tell me they will decide once the technology settles or the guidance is clearer. It will not settle. Readiness is built by deciding in uncertainty and being willing to revise, which is the ordinary condition of leadership, not a special condition of AI.
What good looks like
An AI ready organisation is recognisable by its behaviour, not its inventory. The leadership team can say in one sentence what AI is for, and staff give the same sentence back. There is a short list of approved uses and tools, and everyone knows where it lives and who to ask when it does not cover their situation. Somebody owns AI at leadership level, with a named deputy, and it appears on the leadership agenda on a rhythm rather than in a crisis. When something goes wrong, the response is calm because the route has already been agreed.
Good also looks like honest limits. The best organisations I work with are clear about what they will not use AI for, and say so publicly. That does more for trust among staff, parents and boards than any amount of enthusiasm, because it shows the leadership has looked at the downside and chosen anyway.
And good looks like a rhythm of review. AI ready organisations do not finish. They revisit the three questions every term or quarter, because the answers will have moved.
Where to start
Start by asking the three questions of your own leadership team, then asking them of someone three levels down, and compare the answers. The gap between the two is your readiness gap, and it is more useful than any maturity score because it tells you exactly what to fix. Vague leadership answers mean you need decisions. Clear leadership answers that the front line cannot repeat mean you need communication. Clear answers everywhere except on who owns it when things go wrong mean you need accountability.
Then fix the biggest gap first, say the decision out loud, and put a date in the diary to ask the questions again. That is not glamorous. It is what readiness is.
Working through this with your leadership team
If your leadership team is trying to work out how ready you really are, and what to do about the gaps, this is the kind of work I support through AI readiness reviews and leadership strategy sessions. You can find out more about that work and how I approach it here.
Dan Fitzpatrick is the founder of The AI Educator, a Forbes contributor, a bestselling author on AI in education, and an adviser to school and system leaders on AI strategy. Read more about Dan.
