After-Action Review Template: Agenda, Worksheet and Follow-Up

Short answer

An after-action review is a structured discussion after a defined event, comparing what was expected with what occurred and deciding what to investigate or do next. Use the template below to record evidence, differing accounts, tentative lessons and owned actions. Keep observations separate from explanations, preserve unresolved questions, and set a review date to check actions against observable evidence rather than assume success.

What is an after-action review?

An after-action review (AAR) is an event-bounded learning discussion after a project, operational cycle, campaign, change or incident. Its focus is the gap between expected and actual events, including what went well and what remains unclear. Wharton’s April 2012 guide describes a discussion of intentions, execution, explanations and adaptation, supported by relevant facts and different perspectives.

For this worksheet, keep three boundaries explicit:

• Incident report: records an incident through the applicable reporting process. This AAR does not replace required reporting, investigation or escalation.
• Performance review: addresses an individual’s work and development. This AAR examines a bounded event, not individual ratings.
• Continuous feedback: supports ongoing listening rather than closing one event. Use Continuous Employee Feedback for that broader practice.

The agenda, worksheet and safeguards below are an original proposed practical framework, not a validated intervention or a reproduction of either source’s tool. Use them to prepare a decision, not to certify a cause or guarantee improvement.

Prepare: a checklist before the meeting

Choose a specific event with a clear beginning and end. Avoid a scope such as “communication problems this year.” Prefer “the handovers during last week’s stock transfer.”

Before inviting participants, complete this checklist:

• Intended outcome: identify the plan, standard or commitment that applied at the time. Do not silently substitute today’s expectations.
• Known facts: draft a short timeline and label gaps. Distinguish recorded events from recalled events.
• Relevant evidence: collect available logs, work records, instructions and participant accounts. Note dates, provenance, missing records and access restrictions.
• Participant coverage: include affected upstream and downstream roles, relevant shifts and people who handled exceptions. Arrange an appropriate contribution route for absent participants.
• Facilitator and recorder: name who will keep the discussion on scope and who will maintain the record. Consider another facilitator when the usual leader’s decisions are central to the review.
• Access boundaries: specify who may read the working notes, final record and supporting evidence, and where each will be stored under existing policy.
• Escalation: identify issues that belong in an existing safety, security, grievance or other specialist process. Do not wait for the AAR to address an urgent concern.

Send the scope and prompts in advance. Ask participants to bring both supporting examples and exceptions to their initial explanations.

Run a proposed 45-minute agenda

The timings below are planning suggestions, not a measured optimum. Narrow the scope or schedule another session if important accounts remain unheard.

Start with a brief opportunity for each participant to contribute before open debate. If agreement comes quickly, ask whose experience is missing and what evidence would challenge the apparent consensus.

Copy-ready one-page after-action review template

Copy this compact worksheet into your working document. Keep entries brief and link to authorized evidence rather than embedding sensitive records. Use an attachment for detail that would otherwise overwhelm the page.

AFTER-ACTION REVIEW
Event / project:
Event dates and scope boundaries:
Review date:
Facilitator / recorder:
Participants, roles and shifts represented:
Missing perspectives and how they will be obtained:

ACCESS
Working notes accessible to:
Final record accessible to:
Evidence location and access restrictions:
Retention handled under:
Issues routed separately, without sensitive detail:

EXPECTED RESULT
What should have happened?
Plan / standard / instruction in force:

ACTUAL RESULT
What occurred, and when?
What is recorded versus recalled?
Unknown or unmeasured elements:

SUPPORTING EVIDENCE
Reference | date | what it supports | limitation:

DIFFERING ACCOUNTS
Account A and context:
Account B and context:
Agreement, disagreement and missing coverage:

CONTRIBUTING CONDITIONS
What appeared to help or hinder?
Observation:
Interpretation:
Causal hypothesis requiring a check:

COUNTEREXAMPLES
Where did the suspected condition occur without the problem?
Where did the problem occur without that condition?

LESSONS
Tentative lesson:
Context and limits:
Evidence needed before wider use:

DECISIONS
Test / investigate / retain / stop / no change:
Reason and linked action ID:

UNRESOLVED QUESTIONS
Question | person obtaining evidence | review date:

Attach an action register

Use one entry per decision. “Communicate better” is not specific enough to review. State what will change, where and for how long, or identify the evidence to collect before deciding.

Separate the due date from the review date: completing an action is not the same as assessing its usefulness.

Separate evidence from explanation

Use three labels throughout the record:

• Observation: what a record shows or a participant reports. Attribute an account as an account; do not convert it into a verified event without checking.
• Interpretation: the meaning someone gives that observation, such as “the handover was rushed.”
• Causal hypothesis: a possible mechanism to examine, such as insufficient overlap leaving exceptions unacknowledged.

Do not treat repeated comments as independent confirmation without checking whether they concern separate events or repeat the same account. A changed outcome after an action also does not establish that the action caused the change.

For each hypothesis, ask what would support it, what would weaken it and what is currently missing. If a measurement is unavailable, write “not measured,” not zero. Record uncertainty without forcing a final explanation.

Facilitate without turning the review into blame

Examine instructions, timing, tools, workload and dependencies rather than assigning character labels. Ask “What information was available at that point?” instead of “Why weren’t people more careful?” Keep decision owners visible without using the worksheet to rate individuals.

Preserve disagreement in the final record. Where accounts conflict, capture their context and the next check rather than selecting the most senior person’s version. Invite affected roles and shifts to correct the summary within a stated period; distinguish corrections from unresolved differences.

Do not promise absolute confidentiality. Explain the actual access boundaries and any applicable escalation requirements. If sensitive concerns arise, pause that part of the discussion and use the established route. Keep unnecessary personal detail out of the routine AAR record.

Fictional worked example: checking a handover explanation

Everything in this example is invented for illustration. It is not customer evidence or a measured result.

A distribution team reviews one week of shift handovers. The expected result was that each open exception would have a receiving owner before the outgoing shift finished. The fictional handover log contains entries without receiving-owner acknowledgments.

The initial explanation is that new starters did not understand the handover process. The facilitator records this as a hypothesis, not a finding, and checks the entries against shift coverage.

In the invented records, some unacknowledged entries came from shifts staffed by experienced workers. A shift including a new starter had complete acknowledgments. The accounts also differ: outgoing staff describe leaving written notes, while incoming staff describe not knowing which exceptions still needed acceptance.

These counterexamples weaken the new-starter-only explanation. They do not establish another cause. The group proposes checking whether unclear responsibility during short overlap periods contributed to the missing acknowledgments.

Its fictional action register reads:

• Decision: test a named receiving owner and explicit acknowledgment at each handover for the next operating week.
• Owner: operations coordinator.
• Due date: before the next week’s first handover.
• Dependency: both shift leads confirm coverage and access to the log.
• Review evidence: acknowledgment records, unresolved exceptions, overlap coverage and accounts from both shifts. Record unavailable data as missing.
• Review date: the first working day after the test week.
• Status: agreed, not yet tested.
• Participant update: coordinator shares the bounded test and review invitation with both shifts.

No follow-up outcome is assumed. Even if acknowledgment completeness later changes, the review should examine coverage and other changed conditions before drawing conclusions.

Follow up before sharing a practice more widely

Reconvene on the review date. Check whether the action occurred, inspect the agreed evidence and invite accounts from affected participants. Keep implementation status separate from evidence of usefulness. Revise or stop the action if dependencies fail, conditions change or evidence challenges its rationale. Explain that decision in the participant update.

NASA’s knowledge-management overview emphasizes preserving context when applying knowledge and lists an after-action review resource. It does not validate this worksheet.

If a practice is validated for a defined context, route it into your existing Organizational Learning workflow for review and possible conversion into training. Pass along its evidence, limits and counterexamples, not just the recommendation.

Start by copying the worksheet, choosing one bounded event and naming the people who will review the next decision.