Property Finder cuts incident lifecycle from days to 14 minutes with AWS DevOps Agent

AWS published a customer case study showing Property Finder running autonomous incident management end to end on AWS DevOps Agent. The pipeline starts from an alert and moves through automated root-cause analysis, Jira ticket creation, an on-call page, and finally an auto-remediation pull request, all in about 14 minutes. Previously, detection alone took two to three days. That is a claimed improvement of more than two orders of magnitude on the front half of the incident lifecycle.

Mechanically, DevOps Agent investigates incidents by correlating telemetry and proposes or applies fixes. Property Finder's design keeps a human merge step through pull requests rather than letting the agent push changes directly. That choice matters this week. In a companion post, AWS details how to build a durable audit trail for DevOps Agent by combining the agent journal, Amazon EventBridge lifecycle events and AWS CloudTrail, so security teams can see exactly what the agent did and its impact.

Competitively, autonomous SRE is a hot category. Microsoft's new Copilot Autopilot, PagerDuty and Datadog's AI features, and startups all target it, and AWS's Matt Wood has just launched CloudWatch Omni with evaluators that score agent work quality on live traffic. AWS is building a full loop: agents act, observability grades them, and audit trails prove what happened.

The caveats: a vendor-published case study naturally highlights best-case numbers, and the 14-minute figure measures a pipeline, not the quality of fixes. In a week defined by agent-containment failures, the audit-trail guidance may be the more important half. Watch for data on remediation PR acceptance rates and false-positive root causes.