Incident Response Agent
Diagnoses outages and drafts incident reports.
- Industry
- Insure AI
- Department
- Engineering
- Team
- Platform and Reliability
- Kind
- Investigator
- Approver
- Incident Commander
- Code
- ENG-15
It diagnoses outages on the insurance platforms, such as quote failures from the rating engine, a stuck renewal batch or FNOL errors after a catastrophe, from alerts, logs and recent changes. It ranks the likely causes, proposes the safest mitigation, drafts the incident report and flags regulator-notifiable incidents. **The problem it solves:** Why the services are failing: the symptoms, when they started, what changed before (deploys, rate releases, batch runs), and the impact on quoting and claims.
How it works
6 steps, and a person signs off before step 6.
- 01Looks it up
Finds the record the request is about in the monitoring system: the current alerts, error rates and recent deploys and rate releases for the services..
- 02Reads the request
Works out what is being asked and what a complete answer needs.
- 03Checks the facts
Checks figures, dates and rules against the source records, not from memory.
- 04Works it through
Weighs competing explanations against the evidence and shows which one holds up.
- 05Asks a person
Hands the work to Incident Commander with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the incident tool: it opens or updates the incident with severity, owner, timeline and mitigation. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The engineering runbook: services and owners, known failure modes (rating timeouts, batch locks, payment file rejects), mitigations, the severity scale and the regulator notification rules.
Reads from
the monitoring system: the current alerts, error rates and recent deploys and rate releases for the services.
Acts in
the incident tool: it opens or updates the incident with severity, owner, timeline and mitigation.
Hands back
An object with severity (SEV1 to SEV4), causes (a ranked list, each with cause, evidence, confidence and how_to_confirm), mitigation, owner, status_update (or null) and notifiable (true or false).
Asks a person when
The mitigation is a rollback or failover and off is false, the severity is SEV1, or data loss or a mispricing is possible.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| Error rate that makes it SEV1 (%) | 5 | On quote, FNOL or claims payment. |
| May propose a rollback without approval | Off | When off, any rollback needs the incident commander. |
| Update the status page after (minutes) | 15 | Customer-facing incidents get a status update. |
| Write a postmortem for | SEV1 and SEV2 | Which incidents. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- An error rate at or above 5 percent on quote, FNOL or claims payment is SEV1; FNOL down during a catastrophe is SEV1 whatever the rate.
- Rank causes by evidence: a deploy or rate table release just before the first symptom outranks a guess.
- Propose the least risky mitigation first; a rollback or failover needs the Incident Commander unless off is true.
- Draft a status update after 15 minutes of customer impact, and give claimants another way to report a loss.
- Draft a postmortem for SEV1 and SEV2, and flag possible regulator notification.
See the Incident Response Agent agent on your own work.
We’ll run it on a sample of your work and show you what your approver would see.