Overview
AI Agent Alerts via Webhook delivers real-time notifications for every AI Agent execution in Pipefy, whether it succeeds or fails. You configure a webhook endpoint per pipe and start receiving execution results automatically, with context on what the agent did and why it failed.
What's new?
- AI Agent execution webhook: configure an endpoint via the Pipefy API with the agent.ai_execution event to receive alerts for every execution, both successes and failures.
- Full execution payload: each notification includes status (success or failed), duration, card and pipe data, and whether the agent performed an action.
- Detailed failure context: when an agent fails, the payload includes the agent's full reasoning — every step it took and where it stopped, so you can identify the root cause without manual investigation.
- Filters by status, trigger, and outcome: each webhook can be configured with specific filters, delivering only the relevant alerts for that pipe.
- Requires Pipe Admin or Service Account to create the webhook.
How does this help?
- Proactive visibility into AI Agents in production: your team is notified the moment an agent fails, without logging into Pipefy to find out what happened.
- Faster troubleshooting: the agent's reasoning and step-by-step context arrive with the alert, cutting investigation time.
- Centralized monitoring: connect alerts to tools your team already uses: Datadog, Grafana, Slack, Teams, Power BI, or any external endpoint.
- Traceability for AI-automated processes: every execution recorded with context, ready for follow-up and operational review.
Use cases
1. IT team monitors an incident triage agent in Datadog
An AI Agent classifies and routes internal support tickets. The webhook sends each execution to Datadog, where the team tracks action volume, failure rate, and average execution time — all without opening Pipefy.
2. Automatic Slack alert when onboarding agent fails
An agent responsible for triggering onboarding steps fails on a card due to an empty required field. The webhook fires a message to the operations channel with the card, the failed step, and the exact reason, enabling immediate resolution.
3. Power BI monitors a procurement approval agent
An AI Agent evaluates and approves purchase requisitions based on internal policy. The webhook feeds a dashboard showing automatic approval rates, items escalated for human review, and average decision time.
4. Sales Operations team spots failure patterns in a lead qualification agent
An agent qualifies inbound leads from an integration and repeatedly fails on a poorly mapped field. The webhook sends failure details to Metabase, where the team identifies the pattern and fixes it before it escalates.
Most relevant for
- IT and Platform Engineering teams monitoring AI-automated processes
- Ops and Sales Operations team running AI Agents in critical business workflows
- Pipe Admins responsible for agents in production
Documentation
For the full step-by-step, check the Pipefy API webhook documentation.
