What you will create
An active Automation with one Agent, a Monday schedule, project context, a duplicate-run policy, and one successful manual test.Before you start
You need a working Agent and a repeatable request. Decide:- what should start the work
- what information the Agent needs
- what the result should contain
- who handles failures
- whether the result belongs in Chat or an Issue
Create the weekly summary
- Open Automations and create a custom Automation.
- Name it
Weekly operations summary. - Choose the reporting Agent and the relevant project.
- Write the request, including the reporting period, source information, expected sections, and what to say when data is missing.
- Set the schedule to Monday at 09:00. The current form uses your browser’s local timezone.
- Choose Send to chat. Each real run will create its own conversation.
- Choose what to do when the previous Run is still active. Skip or combine overlapping work when duplicate reports would be harmful.
- Save the Automation.
- Run it manually once, open the linked Chat, and read the real result.
Check the result
The Automation page should show the saved schedule and next run. Run history should link to one Chat containing either the useful summary or a clear error.If no result appears
Open Automation Run history first:- A skipped or combined run should explain that no new Chat was needed.
- A failed run should link to any partial Chat and the Agent Run.
- A timezone mistake should be corrected in the schedule.
- A bad result usually needs clearer instructions, project context, or a different Agent.
