Overview

Investigating poor user experience in RUM typically means switching between session replays, error panels, traces, and performance timelines to piece together what went wrong. AI Investigations automate that first-pass triage. Datadog’s RUM agent inspects the data attached to your views and surfaces ranked, categorized root-cause findings directly in your RUM workflow.

This page lists the available investigation types.

Single-view AI investigation

Run an agentic investigation on a single RUM view to investigate performance issues or identify optimization opportunities on that specific page or screen. Datadog’s RUM agent inspects the view event and its sub-events to identify root causes from sources spanning the application, the backend, third-party libraries, and the user’s network environment.

For more information, see Single-View AI Investigation.

Multi-view AI investigation

Run an agentic investigation across a sample of views that share a slow performance vital. Multi-View AI Investigation extends the same agentic analysis to a population of views, helping you identify what to fix when an entire (view × vital) pair is consistently slow across users. Available from the Optimization page for Loading Time, Largest Contentful Paint, First Contentful Paint, and Interaction to Next Paint.

For more information, see Multi-View AI Investigation.

Operation AI investigation

Run an agentic investigation on a single operation in Operations Monitoring. The agent analyzes both the success rate and the latency of the operation, surfacing focused investigations for each failure mode (errors, timeouts, abandonment) and for latency regressions.

For more information, see Operation AI Investigation.

Further reading

Additional helpful documentation, links, and articles:

Learn about the RUM ExplorerDOCUMENTATION