Sense and decide
Lower the Cost of Speaking Up
Let people rehearse a hard point with AI in private, so more of the room speaks up.
Borrowed from the Cybernetic Teammate field experiment at Procter & Gamble, by Fabrizio Dell'Acqua and colleagues (NBER Working Paper 33641). Read the original source.
Ongoing habit · Minutes to start · Anyone on the team can start it · Start here for this problem
Try this first
Before a hard discussion, tell people they are welcome to think it through with AI first and bring the sharpened point.
Use it when
The same few voices carry every meeting, or good points die unspoken because they arrived half-formed.
Skip it when
This should widen participation, not swap people's thinking for polished model text. Watch for contributions converging on one voice, or people outsourcing thinking they should be doing.
How to introduce it
Invite people to work a contribution out with AI before a discussion: pressure-test a dissent, shape a half-formed idea, find words for a concern they cannot phrase. You are lowering the effort and the risk of speaking, so more of the room does.
How to show up
Say openly that this is welcome, so nobody reads it as cheating or hiding behind a tool. Then make room for what it surfaces. Preparation helps only if the group receives the contribution well.
How long it takes
No session needed. It's an encouraged practice before discussions.
What makes it hard
Some will say prepared points feel less authentic, and some will worry that one shared tool narrows everyone's thinking. Answer the first by valuing the contribution and the second by prizing difference of view. None of it helps if the group does not welcome the voices it draws out.
What it looks like when it's working
More people contribute, junior and reticent members raise substantive points, and the range of what gets said widens. The same few voices, and contributions that all sound alike, mean it is not working.
How long until it sticks
Expect a few discussions before quieter members trust that a prepared contribution will be welcomed.
How you know it stuck
People arrive at hard conversations having thought their contribution through, and the range of voices has widened.
The idea behind it
Good points go unsaid because phrasing them live feels too risky or too hard. A private way to prepare lets in contributions that would otherwise stay silent.
Where it comes from
The Cybernetic Teammate experiment. In the P&G experiment, people working with AI felt better about the work. One reason is that it lowers the bar to contributing, so quieter and more junior people speak. People use AI on their own to sharpen something hard to say. A dissent, a half-formed idea, a worry they cannot yet phrase. Then they bring it to the group.
Why it matters for senior teams
Seniority can make this problem worse because people are less likely to challenge a high-status colleague
Pairs with
Form your own view first and use AI only to sharpen it, not to generate it, if everyone sharpens a point with the same model, people's objections can start to sound alike because they have been shaped by the same model, which defeats the purpose. Say you used it, and speak your own words, not the model's polish.
Evidence
Dell'Acqua et al., The Cybernetic Teammate, NBER Working Paper 33641, P&G field experiment, n=776
Bringing Everyone to the Table: An Experimental Study of LLM-Facilitated Group Decision Making, ACM (2026) In a randomized experiment with 281 groups, an AI facilitator got the quietest members to share more information. It did not improve the groups' final decisions. This move helps more voices get heard, and it should not be presented as a way to improve decisions on its own.