Reference

Scheduled Triggers

Scheduled triggers let you run agents automatically on a recurring schedule or at a specific one-time date and time. When a trigger fires, it creates a job r...

Scheduled Triggersscheduled-triggers / overview

Scheduled triggers let you run agents automatically on a recurring schedule or at a specific one-time date and time. When a trigger fires, it creates a job run that you can review from the Tasks page.

Who can create scheduled triggers

  • Any workspace member who has access to the AI chat and at least one agent can create scheduled triggers.

What you can schedule

  • Daily -- run at the same time every day.
  • Weekly -- run on a specific day and time each week.
  • Monthly -- run on a specific day of the month.
  • One-time -- run once at a specific date and time.

All schedules are timezone-aware. If you do not specify a timezone, the system uses the timezone from your user profile.

What happens when a trigger fires

Each time the schedule is reached, the system starts a new job run using the agent and prompt you configured. The agent executes the prompt with full access to its assigned tools. When the run completes, you can view the result, status, duration, and execution log from the Tasks page.

Run statuses

StatusMeaning
ProcessingThe run is currently executing
SuccessThe run completed without errors
FailedThe run encountered an error and did not complete
PartialThe run completed but some steps encountered issues

Notifications

When a trigger run finishes, you receive a Slack notification if your workspace has Slack connected:

  • Failed runs send an immediate notification.
  • Success and Partial runs are included in a periodic digest.

Slack notifications require both the workspace Slack integration and your personal Slack identity to be connected.

Per-task notification policy

Those defaults are workspace-wide. A task that runs often (for example hourly) can carry its own policy so it does not flood your chat:

  • Ask in chat to update the task's notification policy, or set notificationCadence when scheduling it (immediate, hourly, daily, or off).
  • hourly and daily roll every run in that window into one digest for the task, with run counts and a short summary of what happened.
  • off silences successful runs entirely.
  • Failures are always delivered immediately, regardless of the task's policy.
  • A task with no policy keeps following the workspace-wide defaults above.

Playbook