The Behavioral Autopsy

The behavioral autopsy is the post-failure analysis that answers one question: what behavioral precursors were present before the failure that, if addressed, would have changed the outcome? Almost every AI initiative failure has these precursors. The resistance was visible. The adoption gap was measurable. The authority conflict was documented in meeting notes. The compliance gap was in the risk log.

The purpose of the behavioral autopsy is not to assign blame. I am explicit about this because organizations default to blame when conducting post-failure analysis, and blame produces defensive responses that obscure the actual precursors. The purpose is to identify the behavioral signals that were present and not acted on, so the next initiative does not repeat the same pattern.

1. Timeline reconstruction Build a chronological timeline of the initiative from first proposal to failure event. Include: when stakeholder concerns were raised, when adoption metrics first showed signs of stalling, when technical issues were first reported, when the executive sponsor's attention began to shift, when the change management investment was reduced. The timeline reveals the sequence — and usually shows that the failure was not sudden. It was a gradual process that accelerated.
2. Resistance archaeology Review the meeting notes, email threads, risk logs, and survey data from the initiative for evidence of resistance that was documented and not addressed. The behavioral autopsy almost always finds this evidence. Someone named the concern. The concern was logged. The concern was not resolved. The concern became the failure mode. Identifying this sequence is not comfortable. It is necessary.
3. Adoption gap analysis Pull the adoption metrics from the initiative and trace when the gap between expected and actual adoption first appeared. In most adoption failures, the gap appears within the first three to four weeks and then widens steadily. The inflection point — when the gap appeared — is the moment when intervention would have been least costly. Identifying it retrospectively is instructive for the next initiative: what would monitoring for this inflection point in real time require?
4. Stakeholder behavioral map For each key stakeholder in the failed initiative, reconstruct their behavioral timeline: what they said, what they did, and where the gap between those two things was largest. The stakeholder whose verbal commitment to the initiative was high and whose behavioral engagement was low is the Quiet Blocker. Identifying them retrospectively is the input to identifying them prospectively in the next initiative.