Skip to content

Configure an Autopilot

Turn a tested process into an active automation with a trigger, agent, and reviewable run history.

An Autopilot is a workflow configured to keep operating from a trigger. Create one only after the underlying process and review path have been tested.

AI Agent Autopilot configuration showing its trigger and operating setup.
Treat an Autopilot as an operating process: verify its trigger, owner, access, and review path.

Create the Autopilot

Open Workflows and choose the Autopilot creation option. Configure:

  • a clear name and description;
  • the agent responsible for the work;
  • the execution or output mode;
  • the trigger;
  • an optional runbook describing the expected operating procedure.

Review the displayed trigger summary before saving, especially the schedule and timezone for recurring work.

Review access and approvals

Confirm that the assigned agent has the required connection account and tools. Keep external writes behind an approval step, and leave destructive tools disabled unless the process explicitly requires them.

Run it once manually

Activate the Autopilot, then choose Run now. Inspect the run history and confirm that the steps, outputs, and approvals match the tested workflow.

If your plan’s concurrent-run capacity is already in use, the run may be queued. It is not discarded; open the run view or billing settings to review the limit.

Operate the Autopilot

The Autopilot detail page shows its assigned agent, output mode, trigger, runbook, and recent runs. From there you can:

  • pause or activate it;
  • edit its configuration;
  • start a manual run;
  • inspect run history;
  • archive it when it is no longer needed.

Run now is unavailable while the Autopilot is paused.

Expected result

Continue from here

Navigation