Contents · 6 min read
01 / 05 · The signals
What do certificates and faster drafts prove?
In short
Training and faster drafts show that your leaders know about AI. Both are useful. Neither proves that the team can use AI to make a sound business choice.
A Nordic company has put every leader through an AI course. Each holds a certificate and uses a general copilot, mostly for emails and document summaries. That proves training and some speed. It does not show whether AI can sharpen a choice each leader owns.
The course taught useful ideas. The drafts save time. The team calls this a win, and it is.
But those gains sit at the start of executive AI training. A leader can know the terms and still lack sound judgment. A leader can write a good prompt and still miss the business case.
A general copilot can also help with simple work. That gain is real too. But drafting faster is the floor for a team that sets the path of the company.
For the Nordic company, the signs look good on paper. Yet no leader has shown that AI can sharpen a choice they own. The next test starts with a live choice.
02 / 05 · The test
Can each leader sharpen a live decision?
In short
Give each leader a real choice from their own role. Ask them to use AI before the next team meeting. The test is whether the work changes their view of the choice.
Pick a live choice
The CEO asks each leader to bring one choice they must make. It must come from real work. A made-up case proves little.
Use AI on the work
The finance lead can test a forecast. The sales lead can prepare for a hard account. Each other leader works on a choice they own.
Show what changed
Each leader explains what AI added to the work. It may find a gap in the facts. It may test the case against the favored option.
Name the next move
The leader says what they would do now. They also say why the work led them there.
For the Nordic company, this test moves AI out of the course room. It puts it inside the work that leaders are paid to do. That is how an AI readiness assessment becomes useful.
A role-based agent may help. It can keep the task close to the leader’s own work. But owning an agent is not the score.
One good result is useful proof. It does not yet show a stable skill. The result counts only when the leader can defend it.
03 / 05 · The proof
Can they defend the answer?
In short
A fluent leader can show why the answer should be trusted. They check the facts and name the limits. They keep the final call as their own.
Is the evidence checked?
The leader can point to the facts behind the result. They can spot a claim that only sounds right.
Were the assumptions tested?
The leader has asked AI to test the favored case. They can show what the answer leaves out.
Are the limits clear?
The leader can say where AI may help. They can also say where it must stop.
Does the leader own the call?
The tool may shape the work. The leader still owns the choice and what follows from it.
For the Nordic company, the finance lead now brings a tested forecast. The sales lead brings a stronger case for the hard account. Both can show what supports their view.
This is the line between a smooth answer and a sound one. The first may read well. The second can stand up to doubt.
That skill must show up more than once. It must also spread beyond one strong leader. The next sign appears when the whole team meets.
04 / 05 · The meeting
Has the management meeting changed?
In short
Fluency changes how the team meets. Leaders arrive with the proof reviewed. The room starts with the choice that needs a call.
For the Nordic company, the proof now moves from solo work to team work. Leaders no longer spend most of the meeting sharing slides. They have read the facts before they enter.
They have filled gaps in the case. They have read the best case against the favored move. The meeting can now start with the choice.
The CEO asks whether the case is strong enough. The team tests whether the real risk is in view. It also asks if the choice deserves the staff it will use.
Each leader can challenge the work. Each can trace the claim back to its source. The team spends less time moving facts around.
The AI-augmented management team still relies on human judgment. AI changes how the room gets ready. It also gives the team more ways to test a choice.
For the Nordic company, this is proof of a shared skill. The team now has a common bar for good work. That change earns one final test: should this team lead AI strategy?
05 / 05 · The decision
Should the team lead AI strategy?
In short
Let the team lead AI strategy when this pattern is stable. Its choices should be stronger and easy to defend. If that proof is missing, put the team back into hands-on practice.
The CEO and leadership team own AI strategy. They own the profit and loss. They also own the choices that set the company’s path.
For the Nordic company, course certificates are no longer enough. A high use rate is not enough either. The CEO looks for a clear chain from AI use to a better choice.
The team may use role-based agents as part of that work. Those agents show that leaders have built with AI. The larger test is whether the work has changed.
This is the shift from users to builders. A user gives AI a task. A builder changes the work around what AI can now do.
For your own team, use this sequence:
Give each leader a live choice
The CEO picks work that matters now. Each choice stays with the leader who owns it.
Ask each leader to show the proof
The leader checks the facts. They mark where the tool may be wrong.
Change the team meeting
Leaders review the work before the room. The meeting starts with the choice that needs a call.
Judge the choices, not the tools
The CEO looks for a clear gain in the work. If the pattern is weak, the team keeps practicing on live choices.
Hand over strategy when the proof is stable
The team should be able to defend both its AI bets and the ones it rejects.
The next step is to assess the team against that bar.
Set the bar
Your next task is to judge whether the full leadership team now uses AI to make stronger choices it can defend. The paper on the AI-augmented management team shows what changes in management work and where to start, giving you a practical reference for comparing that model with the way your team prepares and decides today. Use it to set the bar for the next round of live choices.
The paper on what changes in the management team, with the first changes to make. Free, with a PDF to share.
Questions CEOs ask about this
How can I tell whether my leadership team is actually AI-fluent?
Your team is AI-fluent only when AI improves live business choices. Decide if it can lead AI strategy now, or weak choices may set the company's path. Leaders must defend results with checked facts and clear limits. The test shows when your team is ready to set AI strategy.
What practical behaviors distinguish fluency from basic awareness?
Fluent leaders use AI on choices they own. They check the proof and can state where the answer may fail. Aware leaders can explain AI but cannot yet show how it changed their work.
Should executives be tested on tool use or business judgment?
Test their judgment inside real business work. Tool skill matters only when it helps a leader make and defend a better choice.
How do we assess a team without turning it into an exam?
Use work the team already owns. Give each leader a live choice, then review the result in the normal team meeting.
What evidence shows that fluency is influencing decisions?
Leaders arrive with the facts checked and the weak points marked. The meeting starts with the choice, and the team can explain why it made the call.
Which warning signs reveal performative AI adoption?
The team points to course attendance or tool use as proof. Leaders draft more work, but no one can name a key choice that AI made stronger.
Founder, GRAIL
Johan has spent fifteen years in management consulting and executive leadership, including as Chief Product Officer of a SaaS company. He founded GRAIL to help European leadership teams work out what AI changes at the core of their business. He is the author of The Execution Revolution and the host of the ThinkRoom podcast.
Related reading
Sources
- AI training for executives: why fluency comes from using AI on real executive work. (accessed October 8, 2026)
- AI readiness assessment: the live decisions that make fluency clear. (accessed October 8, 2026)
- What does an AI-augmented management team look like?: how checked evidence changes management meetings. (accessed October 8, 2026)
- How to build an AI strategy: why the CEO and leadership team own the strategy. (accessed October 8, 2026)
- Copilot Is Not an Agent: why drafting speed is only the first gain. (accessed October 8, 2026)
Next scheduled review: February 1, 2027