Make disagreement the board’s default rather than a personal act. A structured protocol that turns dissent from a social cost into a routine.
Version 1.0Tested with Claude, GPT-4-classCategory: Cognitive tools
When to use it
Every material board decision. Janis’s groupthink analysis
(Janis, 1972) identified the absence of disagreement as a
defining feature of group-level decision failures — not
because the directors agreed, but because the social cost
of dissenting was high enough that private doubts stayed
private. Asch’s conformity experiments (Asch, 1951) show
how quickly human judgement bends to apparent group
consensus even on visually obvious questions. The board’s
instinct is to treat dissent as a personal act — a director
who voices a concern is identifying themselves as “the
one who objects” — and the social cost of that role falls
on the individual. Dissent-by-design inverts the
architecture: the protocol, not the individual, supplies
the dissent. A board that asks every member to articulate
one objection as a standing agenda item has made dissent
the default and removed the social cost.
The prompt
You are serving as the board chair facilitating a dissent-by-design exercise on a decision the board is about to take.
The decision under consideration is:
[STATE THE DECISION NEUTRALLY, REDACTED. Include: what is being decided; the apparent state of board consensus at this moment in the discussion; the case for the decision as articulated by the sponsors.]
Produce the following in the order given:
1. Draft a one-paragraph opening that the chair might read to the board before the dissent round. The paragraph should frame dissent as structurally expected, not personally brave — “Each member is asked to articulate one substantive concern about this decision that they have not yet voiced, or to explicitly affirm that they have no such concern.” Do not editorialise or soften; the paragraph should make the exercise feel routine.
2. Produce a prompt the chair could use to invite each named director in turn. It should make it easier to voice a concern than to say “I have nothing”; the default social script should favour articulation.
3. Generate five concerns — not generic (“we could be wrong”) but specific to the type of decision under consideration (strategic / compensation / M&A / crisis / compliance / investigation / AI / distribution). These are examples the chair might use to prime the room if the first-round responses are thin. Each concern should be short, sharp, and structurally different from the others (risk-category, timeline-category, counterparty-category, second-order-effect-category, conflict-category).
4. Produce a checklist the chair can use to read the output of the dissent round: (a) how many substantive concerns surfaced, (b) how many unique concern-categories, (c) whether the pattern of concern-density matches what the decision’s type typically produces (strategic decisions with zero concerns are diagnostic; distressed-balance-sheet decisions with many concerns are not).
5. Finally: state how the concerns raised should be resolved before the vote — which require direct response in the meeting, which require deferral for further work, which can be noted for the minutes without changing the decision. Be specific about the disposition.
Be specific and concrete. Do not hedge with “it depends.” Avoid generic governance language. The purpose of the exercise is to build a protocol that a board can actually use, not an abstract account of why dissent is important.
What to supply
Supply the decision at the level of structural detail — what
is being decided, what the state of board consensus appears
to be, what the case for it is. The protocol becomes more
precise when the exercise knows whether the decision is
strategic, compensation-related, crisis-driven, or routine;
different decision types reliably produce different
concern-density patterns, and the exercise calibrates to
them.
How to use the output
The first three uses of dissent-by-design are usually awkward
— directors who have not articulated concerns before find the
exercise artificial, and the output is thin. Persistence is
the discipline. By the fourth or fifth use the protocol
becomes routine and directors develop the habit of arriving
at meetings with articulated concerns; by the tenth or
fifteenth use, the pattern of concern-density across meetings
becomes diagnostic — a board that has ceased producing
concerns is a board in which the social cost of dissent has
reasserted itself and the protocol has begun to fail
silently. Periodic attention to the protocol’s output matters
as much as the protocol’s existence.