Who Owns AI Incident Diagnosis

This fight decides whether AI diagnosis becomes a feature inside the system of record or a new control point above the stack. Datadog is pushing the first model by using the telemetry, topology, on call flows, and case management it already owns to let Bits investigate alerts inside its own product. Traversal is betting that modern incidents spill across logs, traces, code, tickets, feature flags, and chat, so the winning investigator may be the layer that can read all of them together.1 2 3 4 5 6

• Platform owners start with distribution and native data. Datadog says Bits AI SRE is powered by the breadth of its platform data, and its docs show investigations can already use metrics, traces, logs, dashboards, events, RUM, GitHub, and more from inside Datadog workflows.1 2 3
• Independent layers win only if they see something the platform cannot. Traversal is framed around diagnosis across many tools at once, while Rootly is moving in the same direction with read only connectors that pull evidence from observability, code, infrastructure, ticketing, and docs without becoming the data warehouse itself.4 5
• Incident tools and observability tools are converging on the same operator moment. Rootly already uses connected third party systems during incident response, and Datadog lets responders trigger Bits from incident channels, which means the practical competition is over who produces the fastest defensible root cause in the place engineers already work.2 5 6

The next step is a split market. Single vendor shops will default toward the platform that already stores most telemetry, while heterogeneous environments will reward tools that can stitch together evidence across vendors and surface one clear diagnosis. The company that becomes the default first responder will gain the best position to expand into remediation and workflow ownership.1 2 4 5 6