Being busy is no longer a badge of honour
Someone I know who builds AI systems at one of the big tech companies, and with whom I've discussed this topic more than once, shared a reflection in March that struck me as exaggerated the first time: busy people are going to be the biggest losers of the AI era. The more we've discussed it, the more right it seems to me.
The idea in a nutshell
He often talks to leaders at large companies who assure him they use AI "all the time". Dig a little, and using it means asking it to rewrite an email or summarise a PDF. They aren't automating their work, they aren't using it to think, and above all they aren't leading their teams by example. And when he asks why they don't go further, the answer is always the same: I'm too busy.
"If your calendar is a solid block of color from 8:00 AM to 6:00 PM, you are not a visionary leader. You are a bottleneck."
His argument has three parts that I think are key:
- The rules have changed. For decades, being big was an advantage. Now AI is driving the cost of intelligence close to zero, and it's easier for a team of three to build an entire product than for a large company to make a single decision.
- Being busy is a shield. Learning AI forces you to be a beginner again, which is uncomfortable for someone who has been an expert for twenty years. Filling your calendar with meetings is a way of hiding.
- Reading about AI isn't using AI. What makes the difference is practice: trying again and again until you know how to talk to the model and where it fails.
His playbook is blunt: cancel at least 30% of your internal meetings, give your teams ten hours a week to experiment, build prototypes yourself even if they're ugly, and ask for 10x improvements, not 10%.
Why it hits close to home
I've lived in both worlds he describes. On the Tornado and Eurofighter programmes I worked in one of the most regulated organisations there is, where every decision goes through many hands. Bringing digital signatures to NATO's secure network meant convincing the board. It was worth it, because approvals became 60% faster, but I learned what it takes to move something big.
Later, at Laborai, we built an AI-native company from scratch. The pace there was different: we created internal AI tools to take repetitive work off the team so they could focus on the cases that needed a person. The difference between the two worlds wasn't the talent of the people. It was the time each organisation allowed itself to try new things.
That's why the line that made me think most isn't the harshest one, but this: if your team is too busy with their usual work to experiment with AI, you are failing them.
What I think will happen in the coming months
From here on the opinions are mine, not his. These are the trends I see and expect to become visible over the course of 2027:
- Hours recovered will become a metric. Just as companies measure sales or costs today, they will start measuring how many hours each team recovers thanks to AI, and what it does with them.
- Agents will move from demos to daily work. Weekly reports, first drafts, sorting customer requests: many tasks a person does by hand today will be done by an agent, and the person will review.
- We'll see more tiny companies competing with giants. Teams of three to ten people will launch products that used to need an entire department.
- Interviews will ask what you've automated. Knowing how to work with AI will stop being a bonus on your CV and become a given, especially for leadership roles.
- In regulated industries, the brake won't be technology but internal rules. Aviation, banking and healthcare will move more slowly, and the winners will be the organisations that create safe spaces to experiment without putting critical work at risk.
- The gap between those who try and those who nod will become visible. Not between big and small companies, but between people: those who spend real hours working with AI and those who only talk about it in meetings.
What I take away
As a kid I learned how things worked by taking them apart, not by reading about them. AI is the same: you only understand it by using it, getting it wrong and trying again. And that takes exactly what busy people don't have: time.
I'll leave you with the question he suggests asking in your next one-on-one with your team, which I also ask myself: what task that takes you five hours a week today are you going to cut to five minutes this month?
Based on a reflection a contact shared in March 2026 and on the conversations we have had about it.
← Back to the blog