Every call ends the same way: apparatus back in quarters, hose repacked, and a truck full of people who each saw a different fire. The after-action review (AAR) is the one hour that turns those separate memories into a shared account of what happened, and then into training. Most departments know they should hold one. Fewer do it consistently, and fewer still write anything down that a probie will read two years from now.
This guide covers how to run an AAR that people actually want to attend, what to capture, and how to make sure the action items survive contact with the next shift. A copy-and-paste template is at the end.
What an after-action review is (and what it is not)
An AAR is a structured, blameless discussion held soon after an incident to answer four questions:
- What was supposed to happen? The plan, the SOG, the assignment.
- What actually happened? The facts, in order, from the record rather than from memory.
- Why was there a difference? Causes, not culprits.
- What do we sustain, and what do we change? Strengths to keep, improvements to make, and who owns each one.
The format comes from the U.S. Army's Center for Army Lessons Learned, and the fire service adopted it because the problems are the same: high-consequence work, imperfect information, and crews that need to learn from a call without anyone being put on trial for it.
An AAR is not:
- A hot wash. The hot wash is the ten-minute tailboard talk before you leave the scene. It is valuable and you should keep doing it, but it runs on adrenaline and incomplete facts. The AAR is the follow-up, with the radio traffic and the timeline in front of everyone.
- A disciplinary review. If a call is headed for a formal investigation, that is a separate process with different rules. Mixing the two kills candor in every AAR that follows.
- A critique. Older departments used the word, and it carried the baggage. A critique happens to the crew. An AAR happens with them.
Before the review: assemble the record
The quality of an AAR is decided before anyone sits down. Memory reshapes itself within hours; two firefighters who stood on the same porch will honestly report different stretch times by the next day. The fix is to review the incident from its record rather than its recollection.
Hold it within 72 hours. Later than a week and the details are gone. Sooner than the next day and nobody has slept.
Gather, in this order of value:
- Radio audio, with a timeline. Dispatch time stamps and the working-channel traffic settle most "when did we call for the second alarm" arguments in seconds. Most county systems archive it; ask dispatch or check whether your county publishes to a public radio archive.
- CAD times. Dispatched, en route, on scene, and the benchmarks (water on the fire, primary search complete, under control). These are the skeleton of the timeline.
- Apparatus placement and lines. Where the first engine stopped, where the supply came from, and which line went where. A sketch on a whiteboard works; a map with the actual positions works better.
- Photos, ideally labeled by side (A, B, C, D) and time.
- The pre-plan or building information, if there was one, so you can compare the plan to the building you actually found.
Invite the right people. The incident commander, every company officer who had an assignment, and at least one firefighter from each crew. Mutual-aid officers if they had a role. Dispatch is worth inviting for any call where communications were an issue. Keep it under twelve people; above that it turns into a briefing.
Decide who facilitates. It should not be the IC. The IC is a participant with the most to learn and the strongest urge to defend the plan. A training officer or a chief officer who was not on the call is ideal.
Ground rules to say out loud
Say these every time, even to a crew that has heard them fifty times. They are what make the room honest.
- Rank stays at the door. The probie's observation from the front yard counts as much as the chief's from the command post.
- Facts first, then interpretation. Build the timeline before anyone explains anything.
- No names in the improvements. "The second line was delayed" rather than "Engine 3 was slow." The action item can name an owner; the finding should not name a target.
- One conversation. Side conversations are where the real findings go to die.
- What is said here is for training.
Running the review: a 45-minute agenda
Ninety minutes is where AARs go to die. Aim for 45, and never exceed 60.
| Time | Segment | What happens |
|---|
| 0:00 – 0:03 | Ground rules | Facilitator states them. Every time. |
| 0:03 – 0:08 | The plan | IC gives the initial size-up, strategy, and first assignments as they were intended. Two minutes, no defending. |
| 0:08 – 0:23 | The timeline | Walk the incident in order from the record. Play the radio traffic at the moments that matter. Stop at each benchmark and ask "who was where?" This is the segment that runs long; the facilitator keeps it moving. |
| 0:23 – 0:30 | What went well | Go around the room. Every person names one thing. Write them all down; do not debate them. Starting here is not politeness, it is how you find out what to keep doing on purpose. |
| 0:30 – 0:40 | What we would change | Same round. Findings, not blame. The facilitator rewrites anything that names a person into something that names a task or a condition. |
| 0:40 – 0:45 | Action items | For each improvement worth acting on: what changes, who owns it, by when, and which skill area it belongs to. If nobody will own it, it is not an action item; it is a wish. |
On the timeline segment: the single most useful habit is to replay the radio at the decision points rather than summarizing them. Hearing the actual transmission at the moment the second line was ordered ends more arguments than any amount of discussion. If you have the positions on a map, scrub to the same moment and look at where everyone was.
On the going-around-the-room rounds: call on the most junior people first. Once an officer speaks, the room aligns to it.
After the review: the part most departments skip
An AAR that lives in someone's notebook did not happen. Three things need to occur in the week after.
1. Publish it. Write up the summary, the strengths, the improvements, and the action items, and put it where the whole department can read it, including the shifts that were not on the call. The write-up should be short enough to read in five minutes. The attendees list and the facilitator's name go on it.
2. Keep it or lock it. Decide what your record is. Some departments treat the AAR as a training document and keep it internal. Others treat it as part of the incident record. If your state's public-records law (FOIL in New York, FOIA elsewhere) reaches it, know that before you write it, and write accordingly. Once a review is final, lock it: no silent edits six months later. If it needs correcting, the correction is a new entry with a date and a name.
3. Route the action items. Each one needs an owner, a due date, and a status somebody checks. The ones that are really training needs (a drill on the stretch, a hydrant-hookup evolution, an RIT refresher) should be flagged as such and land on the training calendar, tagged by skill area, so the training officer is building drills from real calls rather than from a manual.
Then, at 30 days, ask the one question: did the change happen? An AAR program with no follow-up produces the same three findings every year.
Common mistakes
- Holding it only after the bad ones. A good call has as much to teach, and reviewing it removes the stigma. Aim to review every working fire, every mayday or near-miss, and a sample of routine calls.
- Letting the IC facilitate. See above.
- Debating during the "went well" round. Write it down; argue later if at all.
- Findings that name people. They stop the flow of honest observations for the rest of the year.
- Action items with no owner. Every one of these becomes next year's finding.
- Reviewing from memory. Get the audio.
- No write-up. The review happened; the learning did not.
The template
Copy this into whatever your department uses. It mirrors the fields a review needs at minimum; add sections, but do not remove these four.
AFTER-ACTION REVIEW
Incident: [name / address / date]
Incident type: [structure fire, MVA, wildland, EMS, training]
Conducted by: [facilitator]
Conducted on: [date]
Attendees: [names]
Records status: [draft / published / locked]
1. SUMMARY
What was the plan, what happened, and why the difference.
Two or three short paragraphs, from the record.
2. WHAT WENT WELL
-
-
-
3. WHAT WE WOULD CHANGE
-
-
-
4. ACTION ITEMS
| Action | Owner | Due | Skill area | Status | Training? |
|--------|-------|-----|------------|--------|-----------|
| | | | | open | yes / no |
Skill areas: suppression - water supply - search / rescue -
ventilation - backup line - secondary water - secondary search -
extension - utility control - salvage / overhaul - other
5. LESSONS FOR THE DEPARTMENT (optional)
One or two findings worth keeping beyond this incident, filed by
category: tactics - safety - communication - equipment - other.
Tags: [e.g. mutual aid, mayday, second alarm]
How this works in Coffee Table Tactics
The review above is the workflow the app is built around. The incident record (radio audio, CAD times, apparatus and hoseline positions, photos, floor plan) is already assembled in one place, and the replay plays the radio in sync with the map, so the timeline segment is a scrub bar rather than a debate. The AAR panel has the four fields above, action items carry an owner, due date, status and skill area, and an item flagged for training goes straight to the department's training queue. Reviews move from draft to published to locked, with the name and date on each step, so the record is auditable and nothing is edited silently. Findings worth keeping are saved as lessons learned and searchable by category across every incident.
It is free for fire departments to join.
Frequently asked questions
How long should an after-action review take?
Forty-five minutes is the target; sixty is the ceiling. Longer reviews are usually a sign the timeline was not prepared in advance.
What is the difference between a hot wash and an AAR?
A hot wash is the informal tailboard debrief before leaving the scene. An AAR is the structured, scheduled review held within a few days with the incident record in front of the group. Do both.
Who should run the AAR?
A training officer or a chief officer who was not on the call. Not the incident commander.
Should we do an AAR after every call?
Every working fire, every mayday or near-miss, and a regular sample of routine calls. Reviewing good calls is what keeps the process from feeling like punishment.
Are AAR documents public records?
It depends on your state and on how your department classifies the document.
How do we make sure action items actually get done?
Give every item an owner and a due date, publish the list where the whole department sees it, and review open items at 30 days.