Your organisation doesn't have an AI problem. It has a leadership problem that AI has made impossible to ignore. The companies that win won't be the ones with the most pilots, but the ones that change how they decide, organise and take responsibility.
Every organisation I talk to has an AI strategy. Most also have a stack of Copilot licences, a handful of promising pilots, a steering committee that meets once a month, and a slide titled "Our AI journey" with a road winding pleasantly into the sunset.
What very few have is an organisation that works differently because of AI.
Buying a piano doesn't make you a pianist. Buying licences doesn't make you AI-ready. And yet that's roughly how many AI programmes are run: procure the instrument, book a few lessons, and wait for the concert.
The comfortable illusion
Treating AI as a technology project is wonderfully comfortable. Technology projects have owners, budgets, Gantt charts and, crucially, an end date. You can delegate them. You can put them on a status report in green.
The trouble is that AI doesn't stay inside the technology. The moment it starts to work, it changes who decides what, how work flows, which roles matter and who is accountable when things go wrong. That is not the IT department's job description. That is leadership.
“AI will transcend ‘tech’ talks and be seen as a facilitator for a reimagining of management.”Mark Greeven, Professor of Management Innovation, IMD — AI trends for 2026
Three questions your CIO can't answer for you
Your CIO is probably excellent. But there are three questions no technology function can answer on behalf of the business:
- Which decisions should change? Not "where can we use AI?", which is like asking where you could use electricity. Ask which decisions are slow, expensive or inconsistent today, and what would change if they weren't.
- Who owns the outcome? Not the tool, the outcome. If an AI-supported credit decision goes wrong, "the model did it" is not an answer your regulator, your customer or your board will accept.
- What will we stop doing? The most underrated question in transformation. If AI takes over the weekly reporting, does the weekly reporting meeting quietly survive anyway, now with better charts?
What good leadership looks like here
None of this requires a PhD in machine learning. It requires the things good leaders already know how to do, applied deliberately:
- Own the outcome at the top. One accountable executive per AI-enabled outcome, not one "AI owner" for everything.
- Redesign the work, then add AI. Bolting a model onto a broken process gives you a faster broken process.
- Govern early. Guardrails designed before the build enable scale. Guardrails added afterwards are called "the legal review" and take six months.
- Build literacy where decisions are made. If your executive team's AI education consists of one keynote and a LinkedIn carousel, start there.
One of our AIDIVE training participants summed it up better than I can: “AI transformation is primarily a management and governance task.”
“Stop asking, ‘Which skills do we need for AI?’ and start asking, ‘What becomes our bottleneck once AI succeeds?’” — José Parra Moyano, IMD
So who should lead your AI transformation?
Here's a simple test. Ask around your leadership team who is leading the AI transformation. If the answer is "IT", ask again. If the answer is "everyone", ask a third time. If the answer is a business leader who can name the three decisions they're changing this year, congratulations: you have a leadership question with an owner. That's where real AI transformation starts.
