A decision technique that assigns one person to argue formally against a proposal regardless of personal view, making dissent a legitimate duty and defanging groupthink.
What is the strongest case against this, and who is obliged to make it out loud?
The devil's advocate is a single, assigned role: one person is tasked with arguing against a proposal, formally and regardless of their private view. By making dissent a duty rather than an act of courage, the technique defangs groupthink. The point is not that the contrarian is right; it is that someone is obliged to voice the case against, so the room has to engage with it.
The role is literally ecclesiastical. In the Roman Catholic Church's process for canonising saints, the advocatus diaboli, the devil's advocate, was the official appointed to argue against a candidate's sainthood, marshalling every objectionObjectionGo-To-MarketA common sales objectionView reference → and challenging the evidenceEvidenceValidationData supporting or refuting a hypothesisView reference → for miracles. The office was formalised by Pope Sixtus V in 1587 and existed for nearly four centuries, until Pope John Paul II substantially reduced its role in 1983. The term passed into general use to mean anyone who argues a position they do not necessarily hold, in order to test it.
In decisionDecisionStrategyA recorded decision with context, rationale, and consequencesView reference →-making, the technique is the standard antidote to the failure that the social psychologist Irving Janis named groupthink: the tendency of cohesive groups to converge on a course of action while suppressing doubts, in the interest of harmony. Assigning a devil's advocate re-legitimises the doubt that the group's own dynamics would otherwise silence.
The technique is light to run but depends on getting a few things right:
A worked example. A leadership team is close to agreeing to acquire a smaller competitorCompetitorMarket IntelligenceA competing product or companyView reference →, and the mood is enthusiastic. The head of product is assigned the devil's advocate. She argues the case against: the two codebases share no architecture, so the integration estimate is optimistic; the target's revenue is concentrated in three accounts, any of which could churn post-acquisition; and the founder, central to the product, has signalled he wants to leave. None of these kills the deal, but the team adds an integration discovery phase, a retention clause tied to the key accounts, and a founder transition plan. The decision survives, better than it started.
A devil's advocate is most useful when a group is converging quickly and comfortably on an important decision, especially under time pressure or strong leadership, the conditions where groupthink is most likely. It is cheap, needsNeedUserA user need, pain, desire, or constraintView reference → no preparation, and can be invoked in the moment.
It is less suited to decisions that are already genuinely contested, where dissent is flowing freely and a formal advocate adds nothing. Its failure modes are well known: a half-hearted advocate who offers a weak objection so the group can feel it considered the downside, and the reverse, where assigning the role becomes a ritual that licenses the group to dismiss real concerns as merely the advocate playing a part. The technique works only when the argument is taken seriously on its merits.
The devil's advocate is a reflection framework in the team-process category. Its outputs map to three entity types:
initiativeinsightevidenceRecording the dissent in the graph keeps it attached to the decision it challenged. When the initiativeInitiativeStrategyA large coordinated effort to achieve a strategic goalView reference → is reviewed later, the objections and the counter-evidence are still there, so the team can check which concerns proved real. The case against is preserved as part of the decision's history rather than evaporating when the meeting ends.
Stress-testing a build decision
A team is converging on building a feature in-house when the lead formally assigns one engineer the devil's advocate role for the review. Arguing the case against regardless of personal view, the engineer surfaces a third-party API that covers 80 percent of the scope and a maintenance burden nobody had costed, prompting the team to reopen the build-versus-buy question before committing. Because the objection belongs to the role, the engineer raises it without it reading as obstruction.
Lowering the social cost of dissent
In a team where junior members rarely challenge a senior architect, a pre-mortem review rotates the devil's advocate role to a different person each week. When it falls to a junior engineer, they are expected to argue against the proposed migration, which legitimises a concern about data-loss risk that would otherwise have gone unsaid and forces the room to engage with the strongest counter-case.