Disagree and Commit

In the AI Story

The phrase "disagree and commit" predates Lencioni—it was famously used by Andy Grove at Intel and later by Jeff Bezos at Amazon—but Lencioni's framework specifies the conditions under which the practice actually works rather than degenerating into autocracy. The critical word is "genuinely heard." In a dysfunctional organization, "disagree and commit" becomes a ritual: the leader listens politely to objections, then implements the plan they were going to implement anyway. The team has disagreed, and they are told to commit, but the disagreement did not change the decision, so the commitment is hollow. In a healthy organization, the disagreement genuinely influences the decision—not always by changing the ultimate direction, but by refining it, sharpening it, identifying risks, and integrating insights that the original proposal lacked. The commitment that follows is real because the contributors know their thinking mattered.

AI changes the commitment dynamic in two contradictory directions simultaneously. On one hand, cheap prototyping makes the "disagree" phase richer—any dissenting vision can be built and demonstrated in hours, producing concrete alternatives rather than abstract objections. On the other hand, cheap prototyping makes premature commitment more tempting—the team can try every approach rather than committing to one, producing the paralysis of infinite possibility that Chapter 4 of the Lencioni volume diagnoses. The organizations navigating this successfully have established decision deadlines that force commitment even when the team doesn't feel ready, recognizing that a clear decision made with 70% information and executed with conviction outperforms perfect information that arrives too late to matter.

The practice reveals whether trust exists, because genuine commitment to a decision you argued against is an act of profound team-level vulnerability. You are saying: "I believe this approach is suboptimal, but I trust this team's collective judgment more than my individual certainty, and I will execute as though this were my preferred direction." That statement is performable only in an environment where the team has demonstrated, through repeated experience, that committing to the collective decision rather than optimizing for individual validation produces better outcomes for everyone. Without that experiential foundation, the request to commit feels like the demand to surrender, and the rational response is the hedged commitment that undermines execution—"I'll try this, but I'm keeping my alternative approach alive just in case."

Origin

Lencioni formalized the practice in The Five Dysfunctions of a Team as the antidote to the third dysfunction—lack of commitment. His observation was that teams failed to commit not because they lacked information (the common excuse) but because they lacked clarity about whether a decision had been made. The ambiguity was strategic—it protected team members from the vulnerability of being wrong, because if the decision was never quite final, no individual could be held accountable for it. The disagree-and-commit practice forces finality: the decision is made, the direction is set, and the team's job is now execution rather than continued deliberation.

The AI era test case for disagree and commit is the Trivandrum training Segal documents—twenty engineers simultaneously discovering that their professional identities are being reorganized by a tool most of them had not used a week earlier. The successful navigation of that transition required each engineer to commit to a collective direction ("we are learning this tool together, and we are redefining our roles together") even when individual engineers had profound reservations about whether the direction was wise. The commitment was sustainable only because the trust was deep—the team believed that Segal was present, that the concerns were being heard, and that the collective direction was genuinely aimed at the team's flourishing rather than at abstract productivity metrics. Without that belief, the rational response would have been the flight to the woods that Segal observes among displaced developers nationwide—exit without voice, abandoning the collective experiment rather than committing to its uncertain outcome.

Key Ideas

Commitment requires clarity, not consensus. Teams that wait for everyone to agree never achieve genuine commitment, because consensus on complex decisions is almost never achievable—disagree and commit produces conviction where consensus produces only lowest-common-denominator compliance.

The disagree phase must be real. If the engagement with dissenting perspectives is performative rather than genuine, the commitment that follows is equally performative, producing the appearance of alignment without its substance.

Deadlines force commitment. Teams that wait for certainty before committing never commit, because certainty is not available in complex environments—the discipline of setting decision deadlines and honoring them is what separates commitment from eternal deliberation.

AI's speed penalty. Cheap execution removes the natural pressure that scarcity once imposed—when you can build everything, the forcing function that made teams commit to one thing has vanished, and commitment becomes a discipline rather than a necessity.

Commitment is vulnerability. Agreeing to execute a direction you argued against is an act of trust in the team's collective judgment, performable only when the team has earned that trust through prior demonstrations that collective decisions outperform individual certainty.