# AgentOps

> Operational practices for deploying, monitoring, evaluating, and governing agentic systems.

- Category: AI-Native Product
- Stage: Emerging
- Also called: Agent operations, LLMOps for agents
- Canonical: https://www.joelcsaji.com/ai-dictionary/agentops

## Plain English

DevOps-style discipline for agents that can act.

## Example

A team tracks run traces, tool failures, spend, escalation rates, and safety incidents for deployed agents.

## Why it matters

Production agents need observability and governance beyond normal prompt management.

## Related

- [Trace](https://www.joelcsaji.com/ai-dictionary/trace)
- [Evals](https://www.joelcsaji.com/ai-dictionary/evals)
- [Guardrails](https://www.joelcsaji.com/ai-dictionary/guardrails)

## Sources

- [OWASP Agentic AI Top 10](https://genai.owasp.org/2025/12/09/owasp-top-10-for-agentic-applications-the-benchmark-for-agentic-security-in-the-age-of-autonomous-ai/)
- [Anthropic: Building Effective Agents](https://www.anthropic.com/engineering/building-effective-agents)