← Back to blog

Written by Julie Kongsbak

What does it actually mean to be an AI-mature organisation?

5 October 2026 · By Julie Rønnebæk Kongsbak

There is a lot of talk about AI maturity right now. And we often quickly end up talking about tools, use cases, governance and competence development. Do employees have access to the right tools? Do we have control over what they may use them for? Have we found the good use cases? And have employees received the necessary training?

All of that is relevant. But I have begun to wonder whether it is enough to call an organisation AI-mature. An organisation can have bought Copilot for everyone, run a great many courses, established a central AI unit and have employees who are genuinely skilled at using AI — without necessarily being particularly good at turning all of that into a broader organisational capacity.

Perhaps organisational AI maturity is therefore also about something else: the organisation's ability to continually build experience with AI, learn from it and turn that learning into changes in the way the organisation works.

Individual AI competence is not the same as organisational AI maturity

This is a problem I have become more preoccupied with through my AI leadership coaching. I meet leaders who, over a relatively short time, gain much more experience with AI. They find concrete applications in their own work, become better at assessing when AI creates value, and also begin to notice some of the larger questions about employees' use, quality, responsibility and organisation.

But I also find that this learning does not necessarily travel very far. When I ask how the leaders talk about AI in their leadership team, it often goes quieter. They may know that some of their leadership colleagues use AI a great deal, while others hardly do. But they have not necessarily talked about what each of them is learning, what expectations they have of employees, or what their experience means for the way they lead their areas.

The same can easily happen elsewhere. Some employees become very skilled. A team finds a genuinely good application. IT works with the technology. HR works with competence development. Perhaps the organisation also has an AI network or a central AI unit. There can, in other words, be a great deal of AI activity and a great many competences, without it necessarily becoming shared learning.

If the experience stays with particular people or in particular parts of the organisation, we have first and foremost made individuals and local environments more AI-competent. That does not automatically mean that the organisation has become correspondingly AI-mature.

An organisation needs to be able to absorb new knowledge

This is not actually a new organisational problem. Within organisational research, people have long worked with the concept of absorptive capacity — roughly, the organisation's capacity to absorb new knowledge. The basic idea is that it is not enough to have access to new knowledge. An organisation also needs to be able to recognise the value of it, absorb it and turn it into practice.

I think it is an interesting way to look at AI maturity. AI is developing quickly, and the possible applications are closely tied to concrete work tasks. That is also why I do not think it is possible, from the centre, to decide one's way to all the right applications. Part of the learning has to arise through concrete use. An employee may discover that AI can solve an analysis task in an entirely new way. A leader finds a way of working that frees up time. A team, conversely, discovers a place where AI produces poorer quality or creates a new risk.

The interesting thing is what happens afterwards. Does the experience stay in the employee's own chat? Is it shared with colleagues? Does the leader notice it? Is there someone who assesses whether it is only relevant locally, or whether it ought to change the way more people work? And can the experience from concrete use travel back to those who make decisions about tools, competence development and governance?

I have seen a version of the challenge before

I recognise something of the challenge from my earlier working life in government. In a large agency we worked, among other things, with agile structures that were meant to bring IT and the business closer together and make the organisation better able to develop quickly. There was much that was good in that, but I also experienced how difficult it is to get new ways of working to truly reach all the way around a large organisation. Even when you build new structures to break down organisational divides, new ones can appear between those who are close to the development and those who are not.

I think that experience is worth bringing with us now. A central AI unit can build great expertise and at the same time risk becoming an island of its own. But the opposite is not necessarily better either. If hundreds of employees experiment each on their own, we can instead end up with hundreds of small islands.

What, then, characterises an AI-mature organisation?

I do not think there is one right model. A government agency with several thousand employees should of course not organise itself in the same way as a company of 50. But I think we can say something about what the organisation needs to be able to do, regardless of how it chooses to organise itself.

Employees and leaders need, first, the opportunity to gain real experience with AI. Competences cannot be built through courses alone. The technology has to be worked with on concrete tasks, within frameworks where it is clear what one can and may do. But the experience also has to be able to travel. There need to be places where people share both what works and what does not, and where local experience can become shared learning. And the organisation needs to be able to assess that learning and turn it into practice. A good personal way of working is not necessarily an organisational use case, but some experiences should lead to changed processes, new solutions or new priorities.

At the same time, the movement has to go both ways. The organisation sets frameworks for use through strategy, choice of tools, competence efforts and governance. But the experience from concrete use also has to be able to influence those frameworks in turn. Otherwise the organisation risks having one reality in its strategies and another in actual use.

This is perhaps how I currently best understand organisational AI maturity: not as an end state one arrives at, but as the organisation's ability to continually build experience with AI, learn across the organisation and turn that learning into new practice and new decisions.

What does this mean for you as a leader?

As a leader, you can start in a quite practical way. Ask employees what they actually use AI for — not only whether they use it. Talk about concrete experience in team meetings, and be just as interested in what did not work as in the success stories. And take the conversation into the leadership team: What is each of us learning? What expectations do we have of employees' use? Is there something we are doing locally that others ought to know about?

That may sound banal. But in my conversations with leaders, it is far from a given that those conversations take place.

The individual leader cannot, however, create organisational AI maturity alone. If the experience is truly to travel, the organisation also has to create the connections around the leader. There needs to be a path from local experience to those who work with competences, technology and governance. Someone needs to be able to identify the applications that are worth examining or scaling. And the experience from concrete use needs to be able to find its way back to those who make the larger decisions about the organisation's AI effort.

That is perhaps also why I think we should be careful about only asking how many AI use cases we have, how many people have been on a course, or how widespread the use is.

An at least equally interesting question is:

How good is our organisation actually at learning from the AI we already use?

That is the question I would take into the leadership team.

/Julie