A new Google/Ipsos report confirms what I see every day in my work: most people are using AI — but very few know how to use it well.
EN
DE
NL
40% of workers use AI casually. Only 5% have redesigned significant parts of their work with it. And that 5%? They are 4.5x more likely to have received a raise and 4x more likely to have been promoted.
The gap is not about access. It is about fluency.
Here is what I mean. I recently wrote a complete book in two days — together with Claude AI. Not because I typed “write me a book about AI governance” and hit enter. But because I came prepared: twenty years of domain expertise, original frameworks, real stories, and a very clear brief.
Seven words: “Self-assured, but not arrogant. B2B tone. Expert.”
That is not a prompt. That is self-knowledge translated into direction.
The 95% who are not getting results are treating AI like a new version of Microsoft Word. Command in, output out. Wonder why it sounds generic. The cover letter with “Here is your cover letter” still in the text is not AI’s failure. It is a failure of preparation.
AI fluency is not a technical skill. It is a leadership skill. It requires knowing what you want, knowing your own voice, and knowing how to direct a thinking partner rather than operate a tool.
The report also found that 53% of non-users think AI simply does not apply to their work. That is not a technology problem. That is an imagination problem.
The era of the agentic organization has begun. The question is not whether AI will change your work. It already has. The question is whether you will be in the 5% who shape that change — or the 95% who experience it.
Human Before the Loop — The Executive’s Guide to Governing Agentic AI is available now.
