The leadership team agreed: AI adoption is a priority.
The leadership team agreed: AI adoption is a priority. They bought the tools. They gave the team access. A few people started experimenting and got interesting results. And then nothing happened. Not nothing dramatic. Just nothing next. The experiments stayed experiments. The early wins stayed with the early adopters. Six months later, the company is in roughly the same place, and the most common explanation is “the team just isn’t bought in.”
That’s almost never the real problem. You’ve probably called this a motivation issue, a culture issue, or a hiring issue (“we need an AI person”). It’s a structure issue. The work of making AI operational requires specific functions to be covered. When they’re not explicitly assigned, the work falls between people’s actual jobs and nobody picks it up. Not because they don’t care. Because it’s genuinely not clear whose job it is.
We see this in every company where AI has stalled past initial experimentation. The pattern is the same regardless of industry, team size, or tools. Something is missing that nobody thought to name.
The Protocol
Every successful AI operations effort has five distinct roles present. Not five people. Five functions.
On a small team, two or three people cover all five. On a large team, these might be dedicated positions. The number of people doesn’t matter. What matters is that every role is explicitly assigned and everyone knows who’s wearing which hat.
The Five Roles
Visionary
Sets the direction. Identifies which processes should become playbooks. Pitches AI initiatives. Defines success. Champions rollout.
When it’s missing: Nobody knows what to build. The team works on whatever seems interesting rather than what moves the business. AI projects lack strategic alignment and air cover from leadership.
What breaks without it: Effort without direction. Lots of AI activity, no strategic impact. The team builds things nobody asked for and wonders why leadership isn’t impressed.
AI Operator
Drives the work. Translates direction into playbooks. Runs the build cycle. Collects feedback. Coaches the team through adoption. Holds everything together.
When it’s missing: Projects stall between vision and execution. Pilots get built but never deployed. Playbooks get written but never adopted. The Visionary gets frustrated that nothing is moving. The team is confused about what’s expected.
What breaks without it: The center of gravity is gone. No one owns the process of turning AI experiments into repeatable team workflows. The Visionary tries to fill this role themselves, on top of everything else, and becomes the bottleneck for the entire operation.
Implementor
Configures playbooks in AI tools. Handles the technical build: prompts, automations, integrations. Turns the Operator’s design into a working system.
When it’s missing: Playbooks stay on paper. The process is designed but nobody builds it. Or the Operator tries to do both jobs and the coordination work suffers.
What breaks without it: The gap between “we have a process” and “we have a working system” never closes. Or it closes slowly, because the wrong person is doing the technical work.
One note on this role: in 2026, this is the hat AI itself can wear. Modern tools can configure workflows, write prompts, and handle most implementation when given a clear playbook. For many teams, the Implementor isn’t a person. It’s the AI, directed by the Operator. This means your humans focus on the roles AI can’t fill.
Subject Matter Expert (SME)
Shapes the playbook with domain expertise. Defines what “good” looks like. Validates quality during development. Becomes the first line of support during rollout.
When it’s missing: Playbooks are generic. They work in theory but miss the nuances that only someone who does the work understands. The AI produces output that looks right but isn’t, and nobody catches it until a client does.
What breaks without it: Quality collapses. Or worse, quality becomes a committee decision. Three people reviewing output, each with different standards, conflicting feedback, no final authority. Iteration stalls for weeks because “good” was never defined by one voice.
End User
Uses the playbook daily. Follows the process. Provides real-world feedback on whether it actually works.
When it’s missing: You build something nobody uses. The playbook is technically sound but doesn’t fit real workflows. The people who have to use it weren’t part of building it, so they work around it instead of with it.
What breaks without it: Adoption fails on contact with reality. Feedback comes too late to be useful. By the time you learn the playbook doesn’t fit, rebuilding feels like failure, so the team quietly abandons it.
Dangerous Overlaps
Not all role combinations are safe.
Visionary + Operator (high risk): The leader who sets direction AND drives the build becomes the bottleneck for everything. They can’t delegate because they’re wearing both hats. This is the most common and most damaging overlap, and it’s the default in most companies because nobody else has the mandate.
SME as committee (high risk): Multiple people sharing the quality role without clear authority. Every review becomes a negotiation. Name one SME per playbook with final say.
Operator + Implementor (moderate risk): Workable on small teams, but the Operator who also handles technical configuration tends to default to building and neglect the coordination and coaching work. The technical tasks feel more concrete. The operational tasks feel squishy. Over time, the team loses its driver.
SME + End User (low risk): Often the same person on small teams. Fine. They get both the quality perspective and the practical perspective.
Where This Sits
This is an Operations protocol. It changes how you structure your team for AI work.
The question this protocol teaches you to ask: “For this initiative, who specifically is wearing each hat, and is anyone wearing two that shouldn’t be combined?”
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 →]