Collective Intelligence Is Designed, Not Assembled

Leadership teams are usually assembled for coverage: every function represented, every perspective present. The assumption is that intelligence in the room aggregates automatically — that the group's thinking will be at least as good as its smartest members. Decades of research, and most executives' private experience, say otherwise. Groups routinely underperform the individuals inside them.

The difference between a room full of smart people and a smart room is not talent. It is design: the framing of the question, the distribution of airtime, the way dissent is handled, and the point in the process at which the most senior voice speaks.

Three conditions that change the quality of thinking

First, shared framing. Groups think poorly when members are answering different questions. Making the decision explicit — what exactly is being decided, by whom, against which criteria — is unglamorous work that transforms the conversation that follows.

Second, voice before status. When the most senior view is heard first, every subsequent contribution becomes a reaction to it. Sequencing contributions so evidence and dissent surface early is a structural choice, not a facilitation trick.

Third, recorded reasoning. Groups that capture not just the decision but the logic behind it build a shared model they can revisit and extend. Groups that record only outcomes re-litigate the same questions every quarter.

The payoff

None of this requires new people. It requires treating the quality of collective thinking as something that can be engineered — with the same seriousness an organisation brings to its finances or its technology. When the conditions change, the intelligence already inside the organisation finally becomes available to it.