The AI Coaching Agent runs on a schedule once your coaching policy is set up and activated, and can also be run on demand.* This article covers both ways to run it and how to check on runs.
Before the Agent Can Run
The agent only runs once a coaching policy has been created and activated. If you haven't done this yet, read the article How to Set Up the Coaching Policy.
Automatic Runs
After you save and activate your policy, the agent runs on the cadence you selected during setup (for example, every 2 hours).
What happens on each run
- Reviews new qualifying events since its last run.
- Applies your coaching policy to help determine a recommended action, routing, and escalation for each event.
- Logs each recommendation to the Coaching Agent Log, including the policy clause it applied and its reasoning, so your team can review and act on it.
Run the Agent Manually
You don't need to wait for the next scheduled run. You can trigger a run at any time.

- Go to the Coaching Agent Log.
- Click Run now in the "Coaching Agent is active" banner.
- A confirmation message appears: "Agent run started - decisions will appear here as it works."
Change the Run Cadence
- Go to the Coaching Agent Log.
- Open the coaching policy.
- Update the Run cadence dropdown.
- Click Save & activate to apply the change.
Check on a Run
Use the Coaching Agent Log to confirm the agent is running as expected:
- Runs today shows how many times the agent has run so far today.
- Decisions logged shows the total number of decisions made in the selected date range.
- The log table lists each recommendation with the driver, action, policy clause, and reasoning, so your team can spot-check and confirm the agent is applying your policy as intended.
Click Refresh on the log page to pull the latest runs and decisions. To learn more, read the article How to Set the Refresher Widget.
Turn the Agent Off
- Go to the Coaching Agent Log.
- Click View & edit policy in the "Coaching Agent is active" banner.
- Turn off Activate AI Coaching Agent for this fleet.
- Click Save & activate.
*AI-generated responses may be incomplete or inaccurate - always verify before acting on safety, employment, or operational decisions.