When the CEO Learns to Code: What AI Means for Leadership
Last Friday, my co-founder Jan-Eric built a working prototype of a client proposal generator. Not by hiring a developer. Not by filing a product request. He sat down with Claude Code for two hours on a Friday afternoon and shipped something that would have taken a sprint planning meeting, a Jira board, and three weeks of back-and-forth a year ago.
He's not a developer. He runs an AI training company. But that's exactly the point.
The Shift I'm Seeing in Real Time
The executives I train are different from the ones I trained a year ago. A year ago, they wanted to understand AI conceptually — "what can it do for my team?" Now they want to use it personally. They're not asking for slides about AI strategy. They're asking me to sit with them while they build something.
When I trained tourism executives at CTS, the focus was on strategic thinking — the 80/20 principle of human-AI collaboration. That was mid-2025. The sessions I run now increasingly feature executives who want to prototype their own workflows, not just hear about someone else's.
This changes the training dynamic completely. I used to teach from the front. Now I pair-program with C-suite executives who are debugging their own automations. It's the most interesting my work has been in two years.
The Uncomfortable Part
When the CEO is more AI-fluent than the team, it creates a new kind of friction. I've watched it happen: an executive builds a prototype over the weekend, walks into Monday's meeting, and says "why can't the team do what I did in two hours?" The answer is usually that the team hasn't had the same learning runway. The CEO had permission to experiment. The team has a queue.
This gap is temporary. But it's real and it's frustrating for both sides. When I taught HKJC's management trainees to think in prompts, part of the value was creating a shared language — the same vocabulary for how to talk about AI work, so that a weekend prototype could become a Monday conversation instead of a Monday accusation.
I don't have a clean framework for this yet. The usual advice is "train the team up to the leader's level." That's right but unhelpful. What I'm actually doing in practice is running parallel tracks — a leadership session where the CEO builds, and a team session where the team builds the same thing with more scaffolding. When they compare notes, the CEO gains appreciation for the constraints. The team gains confidence that they can do it too.
What This Means for My Industry
The AI training market is splitting into two tracks, and I don't think most training companies have noticed. There's workforce training — productivity, workflow integration, prompt basics. And there's leadership training — prototyping, decision-making with AI, capability expansion. These need different formats, different depths, different outcomes.
I'm building both simultaneously. Some weeks that feels strategic. Other weeks it feels like I'm spread across too many surfaces. The honest answer is I don't know which track will be more valuable in 18 months. I know which one is more interesting to me right now — watching a non-technical founder build something real in two hours. That moment when they look up from the screen and say "wait, it actually works?" — that's the shift. Everything else is commentary.
I work with leadership teams and organizations on practical AI adoption. The best starting point is usually a conversation about what you've already tried.
