Lizzy Prins shares 4 questions about AI to ask at your next management team meeting

In her debut blog for BusinessWise, technology leadership expert Lizzy Prins warns of a blind spot: boards are throwing themselves into AI training, but missing the strategic vision. Learning to prompt does not make you an AI leader. How can your management team get to grips with AI’s fundamental impact?

The AI literacy requirement has been in force since February 2025, and enforcement can also begin from 2 August 2026. But the question that really leaves your organisation vulnerable is not in the law.

There are many managers around at the moment who feel ready for the future. Take the director who completed a course last month on building AI agents. He now knows how to automate a workflow and connect different tools.

Back at the office, he sees something else. Everyone on his team works with AI in their own way: one person uses it every day, another hardly at all. The managers reporting to him have no idea exactly how their own teams use AI, let alone whether they do so responsibly. An employee cautiously asks whether her role will still exist in a year, and no one can answer the question. Meanwhile, the management team decides: ‘We want to do more with AI.’

Do more with AI. But who decides what that means for whom? Is this a question for IT, HR or him as a manager? No one has the answer, and it is not clear anywhere who is responsible. So despite good intentions, for now nothing structural happens.

Not through unwillingness, but because no one is asking the question that matters most right now: what is actually happening to this organisation as AI seeps into every corner?

That is the problem this article is about. Building an AI agent or learning to write better prompts teaches you nothing about how to answer that question.

Focusing on the tool, blind to the organisation

The AI literacy requirement has been in place since February 2025. In the years since, I have seen most organisations do something: draft some policy or start to establish guidelines. But often without asking the underlying question. What does the AI Act actually mean by literacy? And why would you want this, aside from the law?

Meanwhile, there has been plenty of attention on AI. Managers attended prompt training. Management team members learned about AI agents. Directors tried out the latest tools. As if good leadership begins with becoming the best AI user yourself.

Understandable: starting somewhere is better than standing still. And not unwise either: a leader who has never worked with AI lacks a sense of what the technology can and cannot do. You need that sense to form a considered view of it.

But the higher you rise in an organisation, the less your own skill with AI matters. Not because it is unimportant, but because it answers the wrong question. What a director needs is not a better prompt. It is insight into what is shifting in their organisation.

I believe organisations are confusing two fundamentally different things: AI literacy and AI leadership.

AI literacy is about the relationship between people and AI

AI literacy means people understand what AI is, what it can and cannot do, what risks it involves and how to use AI responsibly in their work.

Can you use AI responsibly?

Can you assess the quality of its output?

Do you know when to use AI, and when not to?

These are important skills, and every organisation should invest in them. But they all concern one relationship: the one between you and the tool. And that is where it ends. Because even if you answer those three questions perfectly for yourself, you still do not know what is happening in your organisation.

AI leadership is about the relationship between AI, people and the organisation

This is where I believe the confusion arises. Many organisations think that managers who can use AI well are automatically prepared for AI’s impact.

But AI leadership calls for a different perspective: not “How do I use AI?”, but the question the director at the beginning of this article had not yet asked himself: “What happens to my organisation once AI becomes part of everyday work?”

That is a very different question, and it cannot be answered with better prompts or by building an AI agent.

Why the CEO does not need to be an AI expert

We expect a finance director to understand the accounts. Not to process every entry themselves. We expect a hospital director to understand what happens in an operating theatre. Not to perform surgery themselves.

With AI, strangely enough, we set the bar in exactly the opposite place. We ask managers to learn how to automate things themselves, while barely discussing what that automation does to their department. We train people in execution and skip over understanding.

Take a customer service team. With AI, most first-line questions can be dealt with quickly. The technology is there, the tool works. But the question that matters is not: ‘Which tool should we choose?’

The question should be: will an employee still be available for difficult, emotional cases? Who decides when a complaint needs a human perspective rather than an automated response? And what happens to the knowledge the team is building now if AI handles most of the conversations in future?

These are questions no prompt training can answer. ## Shadow AI: the organisational risk growing out of sight There is a second example. It is less visible, but just as urgent.

Gartner’s Global Labour Market Survey of more than 12,000 employees and managers shows that 88 per cent of people with access to a business AI tool also use personal AI tools for work. Providing access to a business licence does not solve this. Gartner calls this an enablement illusion: managers confuse access and adoption figures with transformation.

And it is happening largely out of sight. According to the Shadow AI trend report from awareness specialist Awareways, organisations are estimated to have visibility of less than 11 per cent of actual AI use within their own walls.

This does not happen through unwillingness. Often it is out of enthusiasm, or because the work gets done faster. But as soon as business-sensitive information ends up there, it is no longer an individual risk, but an organisational one.

In December 2025, the municipality of Eindhoven announced a data breach: employees had uploaded Youth Act documents, internal reports and CVs to public AI tools. Because of the scale, the municipality could not establish exactly what data was involved. Mayor Dijsselbloem wrote to the council that employees had done this to improve services, not through unwillingness, then, but with good intentions. The enhanced supervision of Eindhoven by the Dutch Data Protection Authority had ended just over two years earlier.

Two years spent working to get privacy policy in order. And still. You cannot solve this by learning to write better prompts yourself or learning to build an AI agent. You solve it by knowing what is happening in your organisation and setting guidelines for it.

These are not technical questions. They are leadership questions.

AI leadership begins in the boardroom, not in ChatGPT

Leadership with AI therefore begins with asking different questions.

So, not:

Which AI tool should we use?

How do I build an AI agent?

Which prompt produces the best output?

But:

• Which work is changing because of AI, and which responsibilities are shifting as a result
• Which decisions will always remain human?
• How do we safeguard quality as AI increasingly analyses, writes or advises
• Which skills will our people need in two years, and who will make sure they have them?

Four questions. Not one is about technology. All four are about choices only you can make. ## Get started straight away: the first steps for your management team Put those four questions on the agenda for your next management team meeting and assign one owner and a date to each. Not IT and HR ‘in consultation’, but one name for each question.

If you do this only because the law requires it, you will get a training course and a certificate, and nothing will change. If you do it because you want to know what is shifting in your organisation, it will give you something no regulator requires of you: control.

Also do one thing you can do today: ask your team leaders to map out this week which AI tools their people actually use and what they use them for. Anonymously, without consequences, and explicitly including personal accounts.

That second question is the most important. A list of tools is an inventory. What people use them for shows which work is shifting. And that immediately gives you material for the discussion with your management team. You will probably be shocked by the answer. That is exactly why you want to know.

The director at the beginning of this article did not need to skip his training on AI agents. But along the way, he should have asked himself a different question. Not: ‘Am I good enough at AI?’ but: ‘Do I know what is shifting in my organisation now that AI is involved in everything?’

That is where AI leadership begins.