Read-only job execution metrics, failure analysis, and process performance queries against Insights, plus alert and access reads.

A coding agent can query Insights for job health, failure trends, and process performance from a plain-language question, without you writing a single query yourself.

The agent answers questions about job execution: success rates, failure counts and their most common causes, processing time, and trends over a period or between two periods. It starts with a summary for context, then drills into whichever process, folder, or time window you care about.

It can also discover the exact folder, process, queue, or machine name you need to scope a question by, and read (but never change) alert definitions, alert history, and a triggered alert's delivery type and recipient count, along with who has Insights access on the tenant.

The agent is read-only in every direction. It doesn't start or stop jobs, create or edit alerts, or change access. Job control routes to uipath-platform, and root-cause analysis to uipath-troubleshoot. Alert edits and Insights access changes are only possible in the Insights UI, and the agent says so instead of attempting them.

This work is powered by the uipath-insights skill, which queries job metrics and failure trends, and reads alert and access-control configuration. You can see it in the Skills catalog.

A first request might look like this:

"Show the processes with the highest failure rates this week."

The agent pulls the job summary for the period, ranks processes by failure count, and surfaces the most common failure reasons. Where you want a rate rather than a count, it divides each process's failures by its total job count.

For the full picture of Insights dashboards and metrics, see the Insights documentation.