The Human Layer: Why AI-First Leaders Still Need to Serve the Room
I spent last week building AI systems in the morning and sitting in rooms full of people in the afternoon. The contrast was not accidental. It taught me something I didn’t expect.
The further I go down the AI-first path, the more clearly I see what AI cannot touch.
The Connecting Thread I Didn’t Plan
On the surface, this week looked scattered. A keynote in Freiburg on AI empowerment for the Mittelstand. A coaching session learning the craft of facilitation next to a master coach. A long wait at Frankfurt Hauptbahnhof watching strangers talk to each other because the trains stopped working. A commitment to serve as a Network Ambassador for a YPO Global Business Summit in Istanbul.
Different rooms. Different contexts. Different languages.
But there was one thread running through all of it: the moments that mattered most were not produced by any tool or system. They were created by human presence, honest conversation, and the willingness to be genuinely useful to the people in front of you.
For anyone building or running a business in 2025 and 2026, that thread deserves more than a passing thought.
AI Is Chefsache — But Not for the Reason Most People Think
In Freiburg, I stood in front of roughly fifty business owners and managing directors and made what I think is still a contrarian argument in the Mittelstand: AI is not an IT project. It is a leadership decision.
Most companies I encounter are treating AI like a software rollout. They assign it to IT or to a motivated middle manager, run a few pilots, and wait to see what happens. The result is a series of impressive demos and very little structural change.
The reason AI needs to be Chefsache is not because the CEO needs to understand every model or API. It is because AI changes roles, changes how decisions get made, and changes what expertise is worth. Those are leadership questions, not technical ones.
More importantly: the unfair advantage in a Mittelstand business rarely lives in the technology. It lives in the niche knowledge accumulated by experienced people over decades — specific customer relationships, process shortcuts, domain intuitions that took years to develop. When that knowledge gets connected cleanly to AI, something more than efficiency emerges. New options become visible. That is a different quality of result than automation alone.
But none of that happens if the CEO treats it as someone else’s job.
What a Broken Train Taught Me About Lowered Expectations
I landed from Bucharest, made it to Frankfurt Hauptbahnhof, and then sat in the DB lounge watching the departure board fill up with cancellations and delays.
What struck me was not the chaos itself. It was how quickly I normalized it. As a local, I had already started calculating buffer time automatically, preparing backup routes, and lowering my expectations before anything went wrong. That habit had become invisible.
And that is the real risk — not the disruption, but the quiet acceptance of it.
I see the same pattern in businesses preparing for a transaction or a major growth phase. Over years, small dysfunctions become routine. Reporting that takes too long. Margin that cannot be fully explained. Key decisions that depend entirely on one person who cannot easily be replaced. Nobody panics, because everyone has adapted. The dysfunction has become background noise.
The moment an investor, a buyer, or a new partner enters the room, background noise becomes foreground risk.
The Frankfurt lounge also reminded me of something else. When the system failed, the room became human. Strangers started talking. Someone made a joke. A traveler asked for help. Suddenly people were present with each other in a way that a functioning departure board never would have created.
Struggle always creates opportunity. Usually not the one we planned for.
The Most Humbling Hour of My Week
I spend most of my working days feeling relatively oriented. I live in conversations about AI agents, local workflows, memory layers, and operating stacks. I build before most people have tested the tool. I am used to moving fast.
Then I sat down to learn facilitation from a master coach in a YPO Game Plan session, and I was a beginner again.
Watching someone hold a room full of successful, opinionated leaders — not by dominating it, but by serving it — was genuinely humbling. The craft was not in the clever question or the sharp observation. It was in the posture. The willingness to give the room your full structure and presence so the group could do its own work, and then let the credit belong to the group.
That is servant leadership in practice. Not as a concept. As a discipline.
And it made me think about what AI can and cannot do in a leadership context.
AI can help me prepare better. It can surface patterns faster, reduce noise, generate options I would have missed, and compress the time between question and answer. I use it for all of that, every day.
But AI cannot manufacture trust in a difficult room. It cannot hold a silence while someone finds the courage to say what they actually think. It cannot replace the credibility that comes from someone knowing you have made the same mistake they are about to make.
Those things require human presence. And human presence requires practice, humility, and the occasional willingness to sit next to someone better than you and take notes.
The Stack That Actually Matters
Here is the mini-framework I have been working with, which I am calling the Leadership Stack for AI-First Operators:
Layer 1 — Tool Layer (Lowest leverage, easiest to copy): The AI tools you use. Models, agents, workflows, automations. Important. Table stakes. Not your moat.
Layer 2 — Knowledge Layer (Medium leverage, harder to copy): The niche expertise in your business, encoded and connected to AI. This is where Mittelstand owners have real advantage — if they choose to use it.
Layer 3 — Human Layer (Highest leverage, impossible to copy): Your presence in the room. Your ability to hold difficult conversations, reduce confusion under pressure, create the conditions where honest thinking happens. This is what AI amplifies but cannot replace.
Most conversations about AI-first leadership focus almost entirely on Layer 1. A few reach Layer 2. Almost nobody talks about Layer 3 — which is, in my experience, where transactions are won or lost, where teams either follow a leader through uncertainty or don’t, and where the real differentiation lives.
What You Can Do Next Week
-
Ask yourself whether AI is Chefsache in your business. Not whether someone is running AI projects, but whether you personally understand how AI is changing roles and decisions inside your company. If you don’t, schedule two hours to find out.
-
Identify one area where lowered expectations have become normalized. Reporting cycles, decision bottlenecks, margin explanations — wherever your team has quietly adapted to dysfunction. Name it. Decide whether it matters.
-
Find your Layer 2. What niche knowledge exists in your business that a competitor or buyer could not easily replicate? Write it down. Then ask how it is currently being captured, shared, or connected to any AI workflow.
-
Enter one room this week as a learner. A peer conversation, a customer visit, a team meeting where you speak last. Practice the posture before you need it under pressure.
-
Separate your tool investments from your human investments. Look at where your time and development budget goes. If Layer 1 is getting all the attention and Layer 3 is getting none, that is a risk worth taking seriously — especially if a transaction or succession is on the horizon.
The week that looked scattered turned out to be coherent. Every room was asking the same question in a different language: how do you stay genuinely useful to the people in front of you, as the world around you changes faster than most of us expected?
I don’t have a complete answer. But I am more convinced than ever that the answer starts with the human layer — and that no amount of AI investment replaces the discipline of serving the room.
What is the one room in your business where your human layer — presence, trust, honest conversation — matters more than any tool you currently use?