The proposal went up on the screen, the person presenting it talked through the rationale, and at the end they asked whether anyone had concerns. There was a pause. A couple of people nodded. Somebody said it looked sensible. The meeting moved on, the decision was recorded as agreed, and everyone went back to their desks.
Three weeks later, when the thing started coming apart, at least half that room said some version of the same sentence: I did wonder about that at the time.
Agreement and the absence of objection are not the same thing
A room can go quiet for a lot of reasons that have nothing to do with being convinced. People go quiet because the person presenting is senior to them, because the decision looks already made, because they cannot articulate the doubt fast enough to say it out loud in front of eleven colleagues, or because the last person who raised something got a slightly cool response and everybody noticed. None of that is captured anywhere. What gets captured is "agreed."
This is the part that makes it expensive. A disagreement voiced in the meeting costs you ten minutes. The same disagreement discovered after the work has started costs you the work.
The tells that a room has agreed to stop talking
There are a few reliable signs, and once you know them you cannot unsee them. The agreement arrives too fast, with no clarifying questions at all — genuine agreement almost always comes with someone wanting to know how a particular edge case gets handled. The comments that do come are about wording, timing or format rather than about the substance. The same two or three people speak, and the rest of the room contributes nods. And when you ask "any concerns?", you get silence rather than someone actually saying "no, I'm comfortable with it," which are very different responses.
Asking "does everyone agree?" is close to useless, because the only socially cheap answer in a room is yes. You are not measuring agreement at that point, you are measuring how hard it is to disagree.
What the silence actually costs
The cost is not the meeting. It is the rework, the second meeting to revisit the decision everyone had supposedly already made, and the slow erosion that happens when people learn their reservations do not change anything. The Australian Financial Review put the price of unproductive meetings at around $39,000 per employee per year, and a decent share of that is meetings held to redo decisions that were never really made the first time.
There is a quieter cost as well. The person with the sharpest read on the risk is often the one least likely to volunteer it — junior, new, or simply not a fast talker. Every time the room takes silence as consent, that person learns something about whether their read is wanted.
Make disagreeing cheaper than staying quiet
The fix is structural, not motivational. Telling people to speak up more does nothing, because the reason they are not speaking up is that the setup makes it costly. Change the setup instead.
Have everyone write their position down before anyone says it out loud, so the first opinion in the room does not anchor the rest. Go round the table and have each person read their own, so nobody can pass invisibly. Ask what would have to be true for this to fail rather than asking whether anyone disagrees — it gives people a legitimate way to raise a doubt without appearing to be the difficult one. And separate the person who owns the decision from the person running the discussion, because those two jobs pull in opposite directions.
All of that is the same principle: get every position onto the table before the room has a chance to converge on the loudest one. If you want to know whether your sessions have that problem, the fit assessment asks ten questions and gives you a straight answer, including a no if that is the honest one.