← Back to blog

Written by Julie Kongsbak

When the leader has to set direction before judgement is built

22 September 2026 · By Julie Rønnebæk Kongsbak

Many leaders are caught in a competence squeeze right now.

On the one hand, expectations are growing. The leader has to set direction for AI, assess the quality of AI-supported work and take responsibility for the decisions in which AI has played a part.

On the other hand, the preconditions often lag behind. Access to good tools may be limited. Practical experience is still limited, and many leaders lack both time and a place to practise.

In the middle stands the leader, expected to set direction before judgement has truly been built.

I use the term the competence squeeze to describe this tension between expectations and preconditions. For me, it is one of the most important challenges in working with AI and leadership right now.

Expectations have grown quickly

AI has in a short time become a leadership task.

Leaders have to consider where AI can create value, how employees use the technology, and which tasks still require human judgement. They need to be able to talk about quality, responsibility, confidentiality, competences and the working environment.

At the same time, they often have to handle large differences among employees. Some already use AI every day. Others are uncertain or hesitant. Some have access to good tools, while others find their own solutions.

The leader must therefore both develop their own use of AI, lead employees' use of it, and navigate organisational frameworks that the leader is rarely in control of themselves.

It is a substantial leadership task. Yet many leaders meet it without much practical experience to stand on.

Uncertainty is not necessarily resistance

When a leader hesitates with AI, it can be tempting to explain it as a lack of interest, technological unease or resistance to change.

But that explanation is not always sufficient.

Sometimes the uncertainty is a perfectly reasonable reaction to a situation where the responsibility is great, while the frameworks are still unclear.

How am I supposed to assess the quality of employees' AI-supported work if I do not have experience with the tool myself?

How am I supposed to be a role model if I am unsure what good use actually looks like?

How am I supposed to set direction if I do not know which tools the organisation will make available in six months?

And how am I supposed to create a shared practice if employees have very different opportunities to use AI?

That is not necessarily an expression of resistance. It can be a sign of a real capacity gap, where expectations have grown faster than the leader's opportunity to build the necessary competences.

Access is necessary, but does not in itself create competence

An approved AI tool is an important precondition. It is hard to develop practical competences without access to something you are actually allowed to use in your work.

But access does not solve the competence squeeze on its own.

A tool on the shelf does not tell the leader how AI can be used in preparing a difficult meeting, in working with a decision, or in developing a work process. Nor does it automatically create shared quality criteria or a language for the dilemmas that arise.

The same applies to a single course. A course can provide knowledge, inspiration and an important shared starting point. But judgement is not developed by hearing about AI alone.

It develops when the leader works concretely with the technology and experiences where it helps, where it falls short, and when the leader themselves need to take over clearly.

Judgement requires time to practise

In my previous blog post on the leader's AI competence profile I wrote that the leader does not necessarily need to be able to build the AI solution. But the leader needs to be able to understand, challenge and prioritise its use. The leader needs to be able to communicate about AI and ensure that the technology is used responsibly.

That requires more than theoretical understanding.

The leader needs to work with AI in their own tasks. That might be preparing a difficult conversation, qualifying a decision, or taking a critical look at a work process.

Through that practical experience, the leader begins to notice the questions that are hard to understand from a distance:

When does an AI response become good enough?

Which parts of the task can the technology support?

Where should a human assess, correct or stop the process?

What consequences does the use have for employees' professional practice, responsibility and way of collaborating?

Judgement arises in the meeting between knowledge, experience and reflection. That is why leaders need practice spaces where they can work with real tasks and at the same time examine their own role.

The competence squeeze must be loosened from both sides

The individual leader cannot solve the competence squeeze alone.

The organisation has a responsibility to create proper preconditions. That includes, among other things, access to relevant tools, clear frameworks, time to learn and the opportunity to share experiences. There is also a need for clarity about which decisions the leader can make themselves, and what should be lifted to, for example, the executive team, IT, HR or legal.

At the same time, the leader must develop their own practice. It becomes difficult to set a meaningful direction for others if AI is still only something the leader has read about or seen demonstrated.

An effort must therefore loosen both sides of the squeeze: make the expectations more manageable and build the necessary preconditions.

If the organisation only formulates ambitions, the leader risks standing alone with the responsibility. If the organisation only offers a tool and a course, it risks AI remaining scattered, individual experiments. What it takes for the organisation to learn from them is something I write about in the post on organisational AI maturity.

The part I can help with

In my 1:1 AI leadership coaching programmes I cannot give the organisation better tools or create clearer organisational frameworks overnight.

But I can help the leader build the practical experience and judgement that make it easier to stand in the situation.

We take the leader's own tasks and dilemmas as our starting point. The leader gets the opportunity to examine where AI can create value in their concrete everyday work, and where there is a need for clear boundaries. At the same time, we work with the leader's view of employees and the organisation.

The goal is not for the leader to know everything about AI. The goal is for the leader to become better able to see the possibilities, ask qualified questions and take responsibility for the direction the organisation is moving in.

AI is changing the leadership role. But if leaders are to be able to take on the new task, they must have the opportunity to develop in step with the expectations.

Otherwise the competence squeeze will only grow tighter.

/Julie