John Laverick

Note

What If the Most Efficient Team Isn't the Best Team?

AI may let smaller teams produce the same or greater output. But output is not the only thing teams produce.

26 June 2026 · Technology & Leadership, AI & Work

One of the most obvious promises of AI is that smaller teams will be able to do more.

That could be genuinely valuable.

Small teams can move quickly. Communication is simpler. Accountability can be clearer. People can spend less time coordinating work and more time doing it.

But I wonder whether we sometimes make the leap from “a smaller team can produce the same output” to “therefore the larger team contained no other value.”

Teams do more than produce.

They are where people learn informally. They are where ideas get challenged before they become decisions. They carry organisational memory. They provide resilience when somebody is away or leaves. They create relationships, belonging and a sense that difficult work is being done with other people rather than alongside a collection of tools.

Those things are much harder to measure than output.

If ten people become five while throughput stays constant, the productivity gain is immediately visible.

It is much harder to see the conversation that no longer happens between somebody experienced and somebody learning. Or the weak assumption that goes unchallenged because there is one fewer perspective in the room. Or the knowledge that becomes concentrated in one person rather than distributed across several.

This isn’t an argument for preserving team size for its own sake.

If AI removes work that does not need to exist, keeping people busy simply to preserve the old structure would make little sense.

The more interesting question is what else the old structure was doing.

If informal learning mattered, how will we create it deliberately?

If peer challenge mattered, where will it come from?

If a wider group created resilience, what replaces that resilience?

If work was an important source of belonging and identity, do we simply treat the loss of that as outside the productivity calculation?

This feels connected to the apprenticeship question, but broader.

Some of what organisations may lose is not junior work. It is the social and organisational capacity created by having groups of people think, learn and solve problems together.

AI may make the production function of a team dramatically more efficient.

That does not automatically make every other function of the team less important.

What if the most efficient team isn’t the best team?

Originally published on Threads.