Aki Kärkkäinen

AI-Empowered Three-Person Teams

Filed under:

The future of teams isn’t 10 people. It’s networks of AI-empowered triads.

Will my job exist in a few years? Focusing on individual roles is understandable, but far less attention is paid to how AI will reshape teams and collaborative work. That’s the more interesting question.

What if the future of high-performing teams is the triad: a small, empowered three-person unit partnered with AI? Whether AI is considered a tool or a team member is an open question—but the distinction matters for accountability and decision-making.

Why Triads?

The Scrum Guide recommends teams of ten or fewer. That was sound advice in a pre-AI world, where a cross-functional team needed enough people to cover all necessary skills. But when AI can cover entire capability gaps on its own, the calculus changes. How many humans does a self-managing team actually need?

Three turns out to be a surprisingly powerful number. In a triad:

  • No one is ever alone with a problem
  • If someone is sick or away, work continues without crisis
  • Cognitive load stays manageable—work gets sliced into small, fast chunks
  • Deep focus time and collaborative flow can coexist, without coordination becoming a bottleneck

But why not dyads (two people) and an AI? Because the third person changes the group dynamics entirely. Dyads tend toward binary tension: agreement or disagreement. A third person introduces mediation, new perspectives, and richer interaction. It also, simply, makes the work more enjoyable.

Think of a music trio: each player is autonomous, but the interplay between three creates something no other combination quite matches. That dynamic tension is exactly what makes the triad powerful.

The Layer Above: Tribal Leadership

A loosely-coupled and distributed network of triads doesn’t operate in a vacuum. With AI handling much of the execution, the real bottleneck shifts to strategy, direction, and culture.

This is where ”tribal leaders” come in. Not managers in the traditional sense, but coaches and connectors who understand team dynamics, embed product thinking, facilitate feedback loops, link teams to customers, and actively eliminate cross-team dependencies. They create the conditions for triads to thrive: shared values, psychological safety, trust, and a clear sense of purpose.

Practices like Troika Consulting—a structured peer-coaching method from Liberating Structures—are exactly the kind of lightweight tools that make this layer work without adding bureaucracy.

Potential weaknesses

No model is perfect. The tribal leader layer creates a potential single point of failure—though this is mitigated by overlapping triad relationships and strong shared values across the network. Onboarding can be harder in a triad: swapping one person changes the entire dynamic, so rotation needs to be intentional.

AI dependency is real: AI tools hallucinate, and small teams may be more vulnerable to over-trusting them without the checks a larger team provides. Not all work suits this model: highly regulated or safety-critical environments may need more oversight than three people can reasonably provide.

Is This Agile For The People Who Gave Up On Agile?

The triad model is lean, human-centered, and built for a world where AI handles capability gaps. It sidesteps the overhead that makes scaled agile frameworks feel like the disease rather than the cure.

If you’ve ever said agile doesn’t work here, this might be worth a second look.

Are you already experimenting with something like this? I’d love to hear what you’re seeing.

Further Reading That Inspired This Article