30 minutes
Five Whys
Follow successive why questions until the group reaches plausible causes it can influence.
A 60-minute AI-facilitated incident review
Blameless Post-Mortem Understand the incident without putting a person on trial.
Investigate causes, organize contributing factors and turn the strongest lessons into owned improvements. Clarvelo keeps the review focused on how the system behaved in context.
3–12 people · Guests join with a link · No workshop board to assemble
Tuesday checkout incident
MI
JO
PR
Five Whys
Theme Sort
Who, What, When
We’re investigating why Tuesday’s checkout incident took 42 minutes to contain. What slowed the response at the moment you were involved? Add what you saw, not a fix.
MI
Mina
Responding to Five Whys
At 14:07 checkout latency crossed four seconds. I saw the alert, but the runbook says to wait five minutes because these spikes usually clear.
Jon
Responding to Five Whys
I held off rolling back at 14:12 because the deploy dashboard still showed the new release as healthy.
JO
PR
Priya
Responding to Five Whys
Support had six failed-payment reports by 14:10, but that channel wasn’t linked in the incident room.
We have three conditions around the slow response: an alert treated as familiar noise, stale deployment data and customer evidence in a separate channel. I’ll put these on the timeline. Before I ask the next why, is there another event or viewpoint this summary misses?
Type your response...
Send
When to use it
When recovery is complete and the team is ready to learn.
Use it after a service incident when a factual timeline exists and the people involved can examine detection, response and recovery without immediate operational pressure.
The aim is system learning, not a tidier account of who made which mistake. Serious regulatory, employment or trust failures may still need independent human expertise.
- Evidence is gathered from several viewpoints
- Symptoms and deeper causes stay distinct
- Important outliers are not tidied away
- Every selected improvement gets an owner
The format
Move from one incident problem to owned system improvements.
15 minutes
Theme Sort
Group causes, contributing factors and action areas without hiding useful outliers.
15 minutes
Who, What, When
Turn the most important lessons into actions with owners and checkpoints.
Your AI facilitator
Protects learning without manufacturing certainty.
Clarvelo collects input from every participant, protects the timeboxes, keeps the outputs connected and explains what happens next.
Teams get reliable process facilitation without a consultant budget for every routine workshop. The diligent organizer gets to contribute, question and vote instead of chairing the room.
Before and after
Arrive prepared; leave with work the team can use.
What to bring
- Confirm that recovery work is complete
- Prepare a factual timeline and relevant evidence
- Invite people from detection, response and recovery
- Explain how the record and actions will be shared
What you take away
- A tested chain of contributing causes
- System themes with uncertainty still visible
- Prioritized improvements rather than general lessons
- Named owners and review dates
Questions
A few useful details before you invite the team.
- What makes a post-mortem blameless?
- The review asks why actions made sense with the information, tools and pressures present at the time. People remain accountable for explaining decisions, while the group looks for conditions the system can improve.
- When should we hold the session?
- Wait until the incident is stable and a basic timeline is available, but hold the review while events are still reasonably fresh.
- Is this suitable for every incident?
- No. Incidents involving legal, regulatory, employment or serious trust concerns may need an independent human reviewer and a different process.
Related workshop formats
Ready when your team is
Start the workshop; leave with a direction the team can use.
Share one link and let Clarvelo guide the group from individual input to a usable output.
Prefer to prepare first? Read our practical guide to running a blameless post-mortem that improves the system .