After-Action Review (AAR): Intent, Actual, Why, Next Time

Teams often finish a launch, a sprint, or an incident and walk into the next piece of work with the same habits. When they do pause, the meeting can split into two unhelpful shapes: a recap with no change, or a hunt for who slipped.

Neither shape produces a usable lesson. You need the original intent on the table, a factual actual, a cause discussion that can survive honesty, and a next-time move someone will actually make. Without those four, “we should communicate better” is only a feeling.

An AAR is a short protocol for that pause. It is not a promise that the room will suddenly feel safe. It is a way to keep the conversation on the work.

An After-Action Review is a structured debrief of one completed event against the intent the team held going in. The method is widely associated with military after-action practice and later civilian project and operations use. Unlike a failure-only post-mortem, an AAR is meant for wins as well as misses. A lucky success can hide a fragile process. Reviewing it is part of the method, not an optional extra.

The four questions run in order: intent, actual, why, next time. “What, not who” is a facilitation rule, not a personality test. If the group cannot name a specific change, it has described the event and stopped short of an AAR.

An AAR is not a performance rating, not a single-cause drill, and not an experiment cycle. It looks backward at one event so the next similar event can be run differently.

A useful AAR is reconstruction plus a change, not a round of opinions.

1. What did we intend to accomplish?

Restate the goal, standard, or plan that was in force at the start. If people disagree about the intent, that disagreement is already a finding. Do not skip this step and jump to complaints.

2. What actually happened?

State the observable outcome. Numbers, timestamps, and who did which step belong here as facts, not as moral scores. Keep spin out. A positive surprise belongs in this step too.

3. Why did it happen that way?

Explain the gap or the match. Look at process, resources, communication, timing, and outside conditions. Multiple causes are allowed. Personality labels (“careless,” “heroic”) are not a substitute for mechanism.

4. What will we do next time?

Name a change: a checklist item, a decision rule, a rehearsal, a handoff, a metric to watch. Assign it if the room can. Without this step, you still have a story.

Run the review close enough to the event that memory is usable. A year-later AAR is a different, weaker artifact.

When another lens fits better, or when you need a complementary view, these frameworks do different jobs. They are not interchangeable labels for the same question.

The After-Action Review is the lens for one completed event: intent, actual, why, next time. Other methods help when the question is a causal chain, a cause diagram, action buckets, a personal loop, a period review, or an experiment.

• After a delivery or sprint. Planning and execution diverged, or they matched for reasons nobody has named yet.
• After an incident. Something broke or nearly broke, and the useful output is prevention, not a person to blame.
• After a high-stakes operation. The next run is soon, and a short, honest reconstruction matters more than a long report.

Example: Feature launch that missed adoption

A product team ships a billing self-serve flow. Intent: 30 percent of eligible customers complete setup in the first week. Actual: 8 percent complete it. Support volume is high.

Implication: The AAR does not start with “whose fault.” Intent and actual are on the table first. Why may include an onboarding email that never sent, a permission error on one plan type, and a help article that described the old flow. Next time: a launch checklist that includes the email, a canary on the failing plan, and a freeze on help-center copy until QA signs it. Dropping the adoption target would be a different decision, not an AAR finding.

Example: A quiet success that should not stay mysterious

An incident commander runs a failover drill. Intent: restore a read-only mode in under 15 minutes. Actual: 11 minutes. The room wants to celebrate and leave.

Implication: Review the win. Why may include a runbook that two people had rehearsed and a dashboard that showed replica lag clearly. Next time: add the rehearsal to onboarding, and do not treat 11 minutes as luck. An AAR on a success is how a fragile win becomes a repeatable move.

What an AAR can help with

• Reconstructing intent versus actual for one completed event
• Explaining a gap without turning the meeting into a rating of people
• Extracting a next-time change from both successes and misses
• Catching the habit of skipping learning because the work “sort of worked”

What an AAR cannot replace

• A single-cause drill. Five Whys follows one causal chain. An AAR can use a why question, but it is a four-step debrief, not a depth-five interrogation.
• A cause diagram. A Fishbone diagram inventories cause categories. An AAR debriefs one event.
• Action buckets without diagnosis. KISS Review sorts Keep / Improve / Start / Stop. It does not require intent versus actual first.
• A personal journal loop. 3R Review (Record, Reflect, Refine) is a lighter individual pass.
• A period-versus-goal writeup. GRAI reviews a stretch of work against a baseline. An AAR is tied to a specific event.
• An experiment cycle. PDCA tests a change forward. An AAR looks backward.
• Psychological safety as a system. The facilitation rule helps. It does not create a safe organization by itself.

Honest scope: AAR structures a debrief and a next-time change. It does not replace the methods above, and it should not be sold as a culture fix.