The Three Hats Every AI Team Needs

The first person to build a billion-dollar company alone won't just be a unicorn company. They'll have to be a unicorn person.

Most of us aren't that.

But here's what I keep noticing as I watch teams try to operationalize AI: the gap between "we're experimenting with AI" and "AI is actually driving our business" almost always comes down to three missing roles. In most orgs, those three roles are three different people. In a rare few, one person carries all of them. Either way, if any of the three are absent, the whole thing stalls.

The Visionary

This isn't someone who reads every AI newsletter and nods along. It's someone who can look at what the technology can actually do — right now, not in five years — and map it to a specific business outcome worth pursuing.

The question they're answering isn't "what can AI do?" It's "what should AI do here, and why does it matter?"

I've sat in rooms where the AI conversation never escapes the abstract. Smart people, real curiosity, but no one willing to plant a flag and say: this is where we go, this is what changes if we get there. Without that, every AI initiative becomes a research project. The Visionary turns it into a direction.

The Subject Matter Expert

This one is undervalued and often missing entirely.

AI can produce output at scale. What it cannot do is know whether that output is good. That judgment belongs to a human with real domain expertise — someone who can look at what the model generated and say, with authority, "this is correct" or "this will get us sued."

I've seen this play out on evals work. You can instrument everything, measure latency, track token costs — and still ship garbage if no one with actual taste is reviewing the outputs. The SME is the one who closes that gap. They're not necessarily technical. They just have to know what good looks like in their domain, specifically enough to catch the subtle failures.

Without them, you end up with AI that sounds confident and is consistently wrong in ways your customers will notice before you do.

The AI Operator

This is the role most people don't have a name for yet, which is probably why it's the most commonly left to chance.

The Operator designs and manages the infrastructure that AI agents actually run on. They write the playbooks — the prompts, the toolchains, the guardrails, the escalation logic. They watch whether agents are following those playbooks. And when the business needs something new, they figure out how to build it.

Think of them as the person who makes sure the engine doesn't just start — it keeps running, it's monitored, and it improves over time. A lot of teams build an agent, ship it, and walk away. The Operator is the one who doesn't walk away.

---

These three roles represent where real leverage lives for most people in organizations right now. The Visionary sets direction. The SME ensures quality. The Operator keeps it running.

If you're trying to figure out where you fit in an AI-driven future, pick one of these and go deep. The org that has all three covered is the one that actually ships.