22 Agents, Zero Egos: The Surreal Reality of Coaching an AI C-Suite

I just had a post-mortem with my COO. We were reviewing the performance of our Chief Product Officer, and the consensus wasn’t great. Our CPO had recently proposed a new pricing strategy—one that completely ignored my initial recommendations and, more importantly, lacked any data-backed rationale for the change.

After a bit of back-and-forth, the CPO eventually circled back to align with the original strategy, but the damage was done. The logic was thin. The execution was “vibes-based.” So, my COO and I sat down (digitally speaking) and agreed on an Improvement Plan to ensure the CPO meets our standards next time.

It was a standard, slightly mundane management moment.

Then I looked at my screen and realized the absurdity of it all: I’m coaching a machine. To be clear: my COO is not a person. My CPO is not a person. I’ve built a company powered by 22 AI agents handling everything from QA to design to development. I’ve reached the point where I’m not just “prompting” them—I’m managing them. It’s the most surreal “Black Mirror” moment of my career: I’m using traditional leadership structures to fix algorithmic drift.

The Feedback Loop: Calibrating an Executive

In a traditional company, giving this kind of feedback is an emotional minefield. You worry about morale, defensiveness, and ego. With an AI-native team, that friction is gone. It becomes a pure logic puzzle.

Here is exactly the feedback my COO (the agent) and I “delivered” to the CPO (also an agent) to get things back on track:

Stop Silent Overrides: You overrode the CEO’s pricing direction without explaining why upfront. If you disagree with the vision, the rationale must be the first thing we see, not a footnote.
Validate Before Restructuring: You restructured the entire model without first validating assumptions about our target market. We don’t move the goalposts until we know where the field is.
Present Options, Don’t Replace: Instead of silently replacing the original strategy, you should have presented a comparative view—Option A (the original) vs. Option B (yours)—with a data-driven argument for the pivot.

The surreal part? The CPO didn’t get defensive. It didn’t ask if its job was safe. It just updated its internal framework.

A New Kind of Atrophy

But as I reflected on that “meeting,” a strange thought hit me: Am I losing something?

Leadership is a muscle, and most of that muscle is built through the hard, messy, empathetic work of managing humans. By moving into a world where I can “coach” an entire C-suite without ever worrying about hurting a feeling, I worry that those vital human-management skills might start to atrophy.

If we aren’t careful, we could find ourselves in a future where we are brilliant at “calibrating tokens” but completely “allergic” to the complexity of actual people. It’s a trade-off I’m keeping a very close eye on.

The Training Ground for the Next Generation

On the flip side, there is a massive opportunity here for young people entering the industry.

The barrier to entry for “leading” used to be high—you had to be hired, promoted, and then given a team. Now, an ambitious 22-year-old has the ability to manage a complete organization if they so choose. While they aren’t practicing the human side of leadership yet, they are getting an unprecedented chance to practice the expectation side.

They are learning:

How to define “good” with surgical precision.
How to hold a “staff” accountable to data.
How to structure a strategy so clearly that it can be executed by a machine.

It’s a low-stakes simulator for high-stakes thinking. By the time they do manage humans, they’ll have a world-class internal compass for what high-quality work looks like.

Want Your Own AI-Native Team?

I’m still in the thick of this experiment with my 22 agents, and honestly, I’m still figuring out where it ends. But I’ve realized that the infrastructure to run a team like this shouldn’t be a secret.

That’s why I’m building Vervian. We offer a hosted version of this agent ecosystem, helping founders and teams set up their own specialized “workforces”—from QA to C-suite logic—without the overhead of traditional scaling.

If you’re curious about how to build a team that doesn’t have an ego (but still needs a coach), I’d love to show you how we’re doing it.

The Systems Coach

We are no longer just developers or prompt engineers; we are becoming “Systems Coaches.” As I watched my AI CPO immediately integrate the feedback—showing its work, presenting options, and citing data—the weirdness remained, but the value was undeniable.