There is a conversation happening in a lot of leadership teams right now, and it tends to follow a familiar shape. Someone raises the question of AI readiness. A few people mention tools they have tried. Someone else mentions a training programme they heard about. The conversation ends with a vague consensus that the team should probably be doing more, and a follow-up that never quite happens.
Nothing is wrong with any of those people. The problem is the question they are trying to answer.
Most businesses, when they think about workforce readiness for AI, are thinking about adoption. How many people are using the tools? How often? Have they completed the onboarding? The assumption underneath all of this is that readiness is a function of exposure: the more your team has touched the technology, the more ready they are.
It is a reasonable assumption. It is also, in practice, almost entirely misleading.
What adoption actually measures
Tool adoption tells you something, but it tells you something narrow. It tells you that people have logged in, run some prompts, maybe saved themselves twenty minutes on a first draft. What it does not tell you is whether any of that has changed the way work actually gets done, or the way decisions actually get made.
And that distinction matters more than most readiness frameworks acknowledge.
Think about what AI is genuinely good at, used well. It is good at compressing the distance between a question and a useful starting point. It is good at surfacing patterns in information that would take a person much longer to see. It is good at holding a large amount of context and returning something structured from it quickly. These are not small things. But they only change outcomes if the person using the tool has the authority, the habit, and the permission to act on what they find.
If your team is using AI to produce faster first drafts but the approval process still takes three weeks, nothing has changed. If someone is using it to synthesise client research but the insight never reaches the person making the strategy call, nothing has changed. The tool has been adopted. The work has not.
The gap that does not show up in dashboards
The readiness gap that most businesses are not solving is not a skills gap in the conventional sense. It is not primarily about people not knowing how to write a prompt or not understanding what a language model can and cannot do. Those things matter, and they are worth addressing. But they are not the ceiling.
The ceiling is structural. It is the gap between what the technology makes possible and what the organisation actually allows to happen.
This shows up in a few specific places. One is decision rights: who is allowed to act on AI-assisted analysis without escalating it for sign-off? Another is workflow design: have any of the actual processes in the business been rebuilt around what AI can do, or has the tool simply been added as an extra step in a process that was designed for a different era? A third is something harder to name, but easy to recognise: the implicit cultural signal about whether using AI is seen as a shortcut or as a skill.
In businesses where AI is genuinely changing how work gets done, those three things have shifted. Not dramatically, not all at once, but enough. People are trusted to use AI-assisted outputs as a real input to decisions. Workflows have been redesigned, even in small ways, to take advantage of what the tools do well. And using AI thoughtfully is understood to be part of doing the job well, not a workaround for doing it less.
In businesses where AI has been adopted but nothing has changed, those three things are still sitting where they were two years ago.
Why the measurement problem persists
It is worth asking why so many businesses default to adoption metrics when they are trying to understand readiness. Part of the answer is that adoption is easy to measure. Seat licences, login rates, usage dashboards: these are concrete, reportable, and they give leadership something to point to.
The harder things to measure are also the things that matter more. Has the quality of decisions improved? Are people bringing better-informed perspectives into the room faster? Is the business learning from what AI surfaces, or just using it to produce output more quickly? These questions do not resolve into a number easily, which means they tend not to get asked in the same systematic way.
There is also a subtler dynamic at play. Measuring adoption puts the responsibility on the individual. If the team is not ready, it is because they have not learned enough, used it enough, engaged enough. Measuring decision quality and workflow design puts the responsibility somewhere else: on the structure of the organisation, on the people who designed the processes, on leadership. That is a harder conversation to start, and a harder one to finish.
A more useful frame
The businesses that are genuinely closing the readiness gap are not the ones running the most training sessions. They are the ones that have asked a different question at the start: not “are our people using AI?” but “has anything actually changed about how we make decisions?”
That question tends to surface the real work quickly. It makes visible the workflows that have not been touched, the approval layers that have not been reconsidered, the cultural signals that are still pointing in the wrong direction. It also, usefully, makes visible the places where things have changed and are working, which is where the model for the rest of the organisation usually lives.
The practical implication is not a new training programme. It is a more honest audit. Pick three or four decisions your business makes regularly, decisions that matter, that take time, that involve synthesising information or forming a view from incomplete data. Ask how those decisions are being made today compared to eighteen months ago. Ask whether AI is in that process in a way that is actually changing the output, or whether it is sitting alongside the process as an optional extra that some people use and some people do not.
The answer to that question is a more accurate picture of where your business actually stands than any adoption dashboard will give you.
Readiness, in the end, is not a property of individuals. It is a property of systems. A team full of people who know how to use every tool available cannot close a gap that the organisation’s own structure is holding open. And a team that has not completed a single formal training session can be genuinely ready if the work around them has been designed to use what they know.
The question worth sitting with is not whether your people are ready. It is whether the organisation they work inside is ready to let them be.
