Scheduling & Automation
Cron jobs, triggers, monitoring. Agents that run without you.
A workflow you trigger manually is useful. A workflow that runs on a schedule, processes results, and notifies you when something needs attention — that's automation. OpenClaw supports scheduled execution through cron jobs, and this is where your agents start running your business instead of just helping with it.
Cron scheduling in OpenClaw uses the standard cron syntax you might know from Unix systems. openclaw cron add creates a scheduled task. You specify when it runs (every morning at 8am, every Monday, every 6 hours) and what it does (trigger a route, run a specific task). The agent executes in the background, writes its results to files, and optionally sends notifications when complete.
Trigger-based execution is the event-driven alternative. Instead of running on a schedule, the agent runs when something happens. A new file appears in a watched directory. A Telegram message arrives. A webhook fires. OpenClaw can be configured to listen for these triggers and kick off workflows automatically. This is particularly powerful for reactive tasks — new lead in CRM triggers enrichment, new support ticket triggers triage.
Monitoring is the piece most people skip, and it's the most important for production reliability. Every automated run should log its results, track success/failure, and alert you on failures. The simplest monitoring pattern: each workflow writes a summary to logs/[workflow-name]/[date].md with status, items processed, errors encountered, and items flagged for manual review. You (or another agent) scan these logs daily.
The automation mindset shift: stop thinking "I'll ask my agent to do this." Start thinking "My agent does this automatically and tells me what I need to know." That's the difference between having an AI assistant and having an AI workforce. The assistant waits for instructions. The workforce runs your operations.
Walkthrough
- 1Choose a workflow to automate — start with something low-risk (daily report, content scan, etc.)
- 2Add a cron schedule:
openclaw cron add --schedule '0 8 * * *' --route [your-workflow] - 3Verify the cron entry:
openclaw cron list - 4Create a logging directory:
mkdir -p logs/[workflow-name] - 5Add logging instructions to your orchestrator: 'Write a run summary to logs/[name]/[date].md with status, items processed, and any errors'
- 6Let the scheduled run execute at least once
- 7Review the log file — does it contain enough information to understand what happened?
- 8Set up a notification: configure Telegram or email alerts for completed runs or failures
- 9Monitor for a week: review logs daily, tune the workflow based on automated results
Checkpoint
You have at least one workflow running on a schedule. It logs its results automatically. You can review logs to understand what happened without watching the agent run.