How to Map top_agent_id to Agent Names in Rovo CSV Exports

[Francisco Castrillo]

When I export Rovo insights as a CSV file, each daily record includes a top_agent_id column. However, I'm unable to identify which agent this ID corresponds to.

Is there a way to map the top_agent_id values to agent names or other identifying information?

Any guidance on resolving these ID references would be helpful.

[Sami Shaik]

Hello @Francisco Castrillo , @Matteo Vecchiato pointed in the right direction, and I tested it in product today, 26 August, on an Enterprise org, because there is one precision that makes or breaks it: the agents export has no column called agent ID. The column you want is "User id."

What is actually going on: Rovo agents surface in your organization's user export as managed accounts. Each agent row carries a User id in the standard Atlassian account format, and the giveaway that a row is an agent rather than a person is its service email ending in @connect.atlassian.com. A masked sample row from today's export:

712020:xxxxxxxx-xxxx-... | AI Learning Coach | <uuid>@connect.atlassian.com | Active | ... | None

So the recipe:

1. Export the user list from Atlassian Administration (the export Matteo referred to).
2. Filter to agent rows using the @connect.atlassian.com email pattern (this also conveniently gives you a full inventory of the agents on your org, including the out-of-the-box ones).
3. Join your insights CSV's top_agent_id against the export's User id column, with a plain XLOOKUP: =XLOOKUP(A2, export!A:A, export!B:B, "not found") where A2 is your top_agent_id and export columns A and B are User id and User name.

One honest boundary: if your top_agent_id values do not visually match the 712020:... account format, compare the UUID portion of the two, and if they still do not line up, reply here with the shape of what you are seeing (masked), because at that point it would be a genuine reporting gap worth filing with Atlassian, and this thread would be a good record of it. Relatedly, https://jira.atlassian.com/browse/AI-1295 already tracks the absence of programmatic access to these same Rovo trends, so the reporting pipeline having rough edges is a known theme.

And if your underlying goal is simply "which agents get used and how much": each agent's owner, managers, and editors can also see per-agent conversation counts in the agent's own Insights, addable as chart widgets on Atlassian Home dashboards, no ID archaeology needed: https://support.atlassian.com/rovo/docs/track-how-often-a-rovo-agent-is-used/

[Kvitka Martsynyshyn]

There's no way to map top_agent_id to an agent name yet — Atlassian hasn't documented one, and there's no public API for it either.

Workarounds:

1. Compare your CSV's daily trigger counts against the Agents tab in Rovo Insights (UI shows names + usage) to infer which ID is which agent.
2. Check each agent's profile URL in Studio/Chat for an embedded ID — may or may not match.
3. If you only have a few unknown IDs, ask Atlassian Support to resolve them from the backend directly.
4. File a feature request — this is a known gap, not something you're missing.

[Matteo Vecchiato]

Hi @Francisco Castrillo ,

Welcome to Atlassian community and thank you for your post.

Have you tried to match the data with the CSV Agents that can be exported from Administration -> Rovo -> Agents ?

Hope it helps