Most companies think AI adoption is a skills problem. It's a structure problem. There's at least one of these roles missing from your org chart, and it's the reason nothing is scaling past the early wins.
There’s a version of AI adoption that looks like it should be working. Leadership is bought in and the tools are there. A few people are using AI daily. Nothing has scaled past them, and you’re stuck in what we call “Random Acts of AI” mode.
Most leaders diagnose this as a “we need to be doing more with AI” problem.
And that instinct isn’t wrong exactly, but it’s strategically useless. Telling thirty people to learn AI is like telling everyone to use Asana (or any project management tool) without any leadership, operational oversight, or people slowing down to figure out what you actually need and build it right.
Every company that has successfully operationalized anything has adopted different hats across the team. With project management, with the CRM, with any system that actually runs.
- There’s the person that sets the vision - why are we doing this and what are we allocating to get what big outcome for the business
- There’s the person who will be the primary stakeholder, they know the nitty gritty of what the solution needs to do in order to be successful
- There’s the person who builds the structure everyone else runs on
- And if that person needs extra technical help, there’s usually a person for that too
With AI, almost nobody has named these different hats.
The 5 hats your team needs to wear.
We work across companies at different stages of AI maturity. Teams just getting started, teams running dozens of AI-assisted processes, and everything in between.
The pattern is consistent: when AI is actually working at a company, five hats are being covered. Not five distinct roles, five hats.
On a small team, two or three people cover all of them. On a larger team, they might be dedicated roles. The number doesn’t matter. What matters is that each function has a name next to it.
We call them hats.
- Everyone on the team needs baseline AI confidence: shared language, basic fluency, the willingness to work alongside AI in their normal job. That’s the foundation.
- Then there are subject matter experts who define what “good” looks like for specific processes.
- There’s a technical function, the implementation, which increasingly AI itself can handle.
- There’s the visionary who sets direction and decides where AI creates real leverage.
And then there’s the one most companies are missing — the AI Operator.
The AI Operator is usually what you’re missing.
This is the hat worth understanding.
The AI Operator translates what the subject matter expert knows into something repeatable: documented playbooks, clear handoffs between humans and AI, guardrails so things fail loudly instead of quietly.
They coach the team through adoption and become the center of gravity for the entire AI operation.
This is not a technical role. Companies keep hiring “AI people,” meaning someone who knows the technology, when what they actually need is someone who thinks in systems and can drive change.
The Operator is closer to an operations manager than to an engineer. They already think in workflows. They already translate between departments. They already hold things together. The difference is they now do it with AI playbooks instead of spreadsheets and SOPs.
Most companies have a visionary: the CEO, the founder, the leader who said “we’re doing this.” And they have everyone on the team using tools at varying levels.
But between the person setting strategy and the people doing daily work, there’s a gap. Nobody is building the bridge.
The visionary can’t do it because they don’t have time for the execution weeds. The team can’t do it because nobody gave them the framework or the mandate. So the work sits in between, undone, and everyone blames the technology.
The most common version of this is the CEO who’s building automations at midnight. They’re setting direction and trying to drive execution at the same time, and they don’t have capacity for both.
The team waits for them on everything. When the CEO gets pulled into something else, AI progress stops entirely.
That’s not a sign that AI is hard. That’s a sign the Operator hat is sitting on the shelf with nobody wearing it.
For leaders: don’t rush into hiring, look internally first.
Here’s what’s interesting about this: the people you need are probably already on your team.
Think about who informally holds processes together right now. The person who designed how the team actually uses your project management tool. The person other people go to when a workflow breaks. The person who thinks in systems, not tasks.
AI skills layer on top of instincts and experience people already have.
We’ve seen operations leads, chiefs of staff, senior managers step into the AI Operator role and run with it.
We’ve seen creative directors, VPs of sales, lead product managers all jump into the Subject Matter Expert role and thrive.
What the visionary actually needs to do here is what visionaries are already good at: set the direction, find the right people, and give them the space to build.
When this clicks for a leadership team, the conversation is almost always the same. They stop asking “how do we get everyone using AI?” and start asking “who’s our Operator? I think it might be Sarah. who’s our Subject Matter Experts on the team? Johnny for creative, Megan for sales, and we probably need someone stronger for marketing…”
That shift, from a vague skills problem to a specific structural decision, is where things start moving.
Start now.
The reason this matters now is that “everyone needs to learn AI” has become the default advice, and it’s creating a specific kind of paralysis for leaders.
It’s technically true. Everyone does need some level of AI competency.
But without differentiated expectations, it means leaders are worried about getting thirty people really good at something that changes every week. That’s overwhelming and it’s the wrong frame.
The hats framework makes it concrete.
Most of your team needs baseline confidence: knowing how to work with AI, comfort with writing and using basic playbooks, exercising basic judgment on outputs, flagging problems early. That’s achievable.
Then, start investing in a few people to develop deeper skills: the subject matter experts who define quality, the Operator who builds systems. And the visionary needs to lead the people wearing the hats, not try to wear all of them.
This is a Foundational protocol. It changes how you structure your team for AI work.
The question it teaches you to ask: “Who is our AI operator? What hat are we missing for this project?”
Know someone dealing with this? If you’ve got a colleague whose company bought the AI tools, did the training, and still can’t figure out why nothing scaled, send them this. The answer is almost always the same: nobody’s wearing the right hats… especially the AI Operator.
Want to map your team? We built a leadership exercise that helps you identify who’s wearing which hat and where you’re thin. It takes 15 minutes and works best when your leadership team does it together. [Get the Five AI Ops Hats Guide]
Read next: Protocol: Own the Playbook, Rent the Tech
Figure out your path: Who Is This For?
Stay connected:
I lead a team and want to build an AI-capable organization. [Subscribe to Leading Momentum]
I’m operationally minded and want to build AI operations skills. [Subscribe to the Operator newsletter]