Workspace reports
Generate a read-only AI report across visible clients, projects, tasks, and notes, with grounded evidence and report history.
- Who can use this
- Members with agent runtime access can read reports
- Generating a report needs a plan with AI access and either hosted AI credits or a valid, explicitly selected OpenAI or Anthropic key
Explore sequence diagram
Sequence diagram
Workspace reports
Sequence diagram. The text on this page has the exact steps and permissions.
Overview
A workspace report summarizes client follow-up, delivery risk, task pressure, and note context. It is read-only: generating a report never changes workspace records.
What a report contains
- A summary of the current workspace state.
- Priority groups, risks, and suggested next actions.
- Grounded evidence rows that cite the visible records used.
- A link to the run trace in Agent Activity.
Example
A Monday report across clients
You generate a report before your weekly planning. It groups Northstar's overdue tasks under delivery risk, lists two requests waiting on a reply, and suggests chasing the open rollback decision. Each point cites the records it used, so you can open them and check before you act.
Scope
Reports only use records you can see: clients, projects, open and overdue tasks, notes, and requests needing a response. When you are working in a team, the report uses that team context.
History
Earlier reports stay listed with who generated them, when, and how many evidence rows and groups they cited. Evidence counts describe the records each report cited, not total workspace coverage.
FAQ
Does a report change any records?
No. Workspace reports are read-only summaries recorded as agent runs.