Event-driven agent
An event-driven agent is an LLM agent triggered by external events (webhook, queue, schedule, system signal) rather than direct chat — handles tickets, monitors logs, sends reminders, runs ETL with reasoning.
Chat agents wait for human input; event-driven agents react to system events. Examples: webhook fires when GitHub PR opens → agent reviews it; Stripe webhook fires on failed payment → agent drafts apology email; CloudWatch alarm fires → agent triages logs and pages a human if needed. Architecturally event-driven agents pair an [[workflow-trigger]] (webhook, queue, cron) with a durable executor ([[workflow-engine]], [[background-agent]]) and an LLM reasoning step. Production patterns: cap cost per event ($X budget cap), retry on transient failures, escalate to humans on confidence drops, dead-letter for repeated failures. Most production AI 'automation' falls into this category — not chat, not autonomous agents, but event-triggered LLM execution.
When to use event-driven agent
- System integrations where reasoning beats hard-coded logic.
- Asynchronous ticket triage, log monitoring.
Common mistakes
- No cost cap — runaway loop burns budget.
- No escalation path — agent silently mishandles edge cases.
FAQ
What is event-driven agent?
An event-driven agent is an LLM agent triggered by external events (webhook, queue, schedule, system signal) rather than direct chat — handles tickets, monitors logs, sends reminders, runs ETL with reasoning.
When should I use event-driven agent?
System integrations where reasoning beats hard-coded logic. Asynchronous ticket triage, log monitoring.
What are the most common mistakes with event-driven agent?
No cost cap — runaway loop burns budget. No escalation path — agent silently mishandles edge cases.
Related terms
- Workflow trigger — A workflow trigger is the event that starts an automation run — webhook, schedule, app event (new Stripe charge, Slack message), or instant push from a partner platform. The right trigger choice determines workflow latency and cost.
- Background agent — A background agent is an LLM-driven worker that runs asynchronously — receives a task, executes for minutes/hours without a user attached, posts results when done. Cursor's Background Agents, Claude Code's async tasks, Devin are 2026 examples.
- Approval workflow — An approval workflow is the agent pattern where high-impact actions (send email, make purchase, delete data) pause for human approval before executing — the production-safe alternative to fully autonomous agents.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/event-driven-agent.md.