Rootly AI SRE and Operate: compared, and used together
Updated 7 October 2026
Rootly AI SRE is part of Rootly's incident management platform. Rootly describes it as "an AI-powered investigation and response engine built directly into Rootly's incident management platform" that requires "explicit human sign-off" before any change (rootly.com).
Operate is a separate, self-hosted AI SRE, and Rootly is listed among Operate's incident-management integrations. So this is less a replacement decision than a question of where the investigation should run, and which data it should read.
Rootly AI SRE vs Operate at a glance
| Operate | Rootly AI SRE | |
|---|---|---|
| What it is | A standalone, self-hosted AI SRE | An investigation engine "built directly into Rootly's incident management platform" |
| What it needs | Docker in your infrastructure, plus read-only credentials | Rootly's platform: it is "natively integrated with Rootly On-Call, Incident Response, Catalog" |
| Code access | Read-only, including repositories | "has read access your code repositories" (Rootly's wording) |
| Human approval | Always. The output is a patch file an engineer applies | "every change requires explicit human sign-off before execution" |
| Model training | Only a billing usage record is sent to Operate | "Rootly enforces zero third-party model training" |
| Where it runs | A Docker image in your own infrastructure. No new cloud accounts, no SaaS endpoints | Not stated on the AI SRE page |
How Rootly AI SRE works, in its own words
Rootly's AI SRE page says it "analyzes your code changes, telemetry, and past incidents to quickly identify root causes and the fix". It starts when an alert fires, runs hypothesis checks in parallel, and gives each finding a confidence score.
It "investigates, surfaces findings, suggests remediation steps, drafts updates and fixes", and Rootly says it "never auto-remediates without human sign-off". Because it lives inside Rootly, it already knows service ownership, on-call schedules and incident history.
How Operate works
Operate is the self-hosted AI SRE we build. Four AI agents read your logs, database and code to work out what broke and why, and a second AI model checks every answer before anyone sees it.
It runs as a Docker image alongside your app, with read-only access to everything it connects to. It watches alerts and errors on its own, and anyone on the team can ask it a question from Slack, MS Teams or another chat tool. Every investigation ends at a .patch file. An engineer reviews the evidence, applies the patch and decides what ships. See how Operate keeps your environment safe.
When Rootly AI SRE is enough
If you already run Rootly for on-call and incidents, its built-in AI SRE keeps investigation in the same tool, with service ownership from Rootly's Catalog and no separate deployment.
When to add Operate alongside Rootly
- The investigation has to run inside your infrastructure, with your own or a self-hosted model.
- You want root causes for problems that never become incidents: a background job that fails quietly, a page that slows down week by week, a support ticket. Product and support can ask from chat.
- You want the fix as a patch file a reviewer applies, never a change executed by a tool.
Operate doesn't do on-call scheduling or incident management. It is not a replacement for Rootly's platform.
Sources
Checked on 7 October 2026.