When Part of the Team Isn’t Human
By Jeff Chung - Chief Strategy Officer
07/01/2026
Helping is one shape AI can take inside a company. Participating is another.
The first shape is what almost every AI deployment looks like today. A person at the center of the work, AI one step removed, helping the person move faster. The work moves through the same chain at a higher rate, and a human still makes the call at every step.
The second shape is the one many teams aren't built around yet. The AI handles a slice of the work end to end, and the human's role shifts to designing the work rather than doing it. The loop closes inside the system, and a person decides what kind of loop should have closed.
That second shape is what changes everything about how the team is organized.
What Has to Change When AI Participates
When AI helps, the work structure barely has to move. The human is still the worker, still the decision point, still the place where the loop closes. AI is fast inside the existing shape, which is why deployments feel safe and outcomes feel marginal.
When AI participates, the structure stops being optional. There is now a part of the team that does a step of the work without a human standing at the center of that step. Ownership has to be defined. Monitoring has to be designed. Escalation paths have to be explicit, because there is no longer an implicit human pulling the work back together if something drifts.
The discomfort organizations feel here is real. Responsibility implies consequence. Avoiding that conversation keeps AI confined to assistance, which is exactly why pilots plateau without producing the kind of impact that justified them.
You Are Standing Up a Role, Not Deploying a Tool
The language people use to describe AI projects still treats it as deployment.
Stand up the tool. Pilot the platform. Roll it out.
The work that actually has to happen at this stage looks more like onboarding. Define the role. Set the scope. Decide what the role is authorized to do on its own, and where it has to pause and wait. Choose how its performance will be observed. Decide where it sits in the team.
This is the verb shift that matters. You are not deploying an AI. You are standing up a specialist. The work is closer to staffing a role than to picking a vendor, even though the system on the other end is software.
The companies treating it as deployment keep getting deployment-shaped outcomes. The ones treating it as workforce design start getting workforce-shaped leverage.
The Structure Is Not Technical. It Is Organizational.
What's interesting about this shift is that the technology is already past the bar. The models are capable enough to participate in real work today, in real ways, across real domains. What hasn't moved at the same pace is the organizational design around them.
Where does this specialist's authority end. Who notices when it's stuck. What happens when its output is outside the bounds of what the team expected. Where does work flow back to a human, and where doesn't it. These are not technical questions. They are organizational design questions, and many of them have not been answered yet because the structure underneath was carrying the work informally.
A specialist that holds a slice of work surfaces every place that informal structure used to live. The technology is ready. The design has to catch up.
Hybrid Teams By Design, Not Accident
Hybrid teams are already showing up in pockets across industries. The question is not whether they form. They form on their own once organizations start standing up specialists. The question is whether they are designed for or whether they accumulate without anyone calling the shape of them.
Organizations that design for it tend to gain real leverage. The structure holds, the work flows, and the specialists do the work they were stood up to do. Organizations that don't tend to accumulate friction. The structure doesn't hold, the work routes around the specialist, and the leverage that was supposed to show up doesn't.
The future shape of work is not human versus AI. It is human and AI inside a structure that was designed for both of them. That structure does not arrive by accident. It is something the organization has to choose to build.
Chief Strategy Officer, Arkane Digital
Jeff advises organizations on AI transformation, focusing on connecting business strategy to practical implementation without fragmented or reactive adoption.