A design where a person reviews or approves an AI decision before it takes effect.
It is a strong control for high-stakes actions, and a weak one for scale. When an agent handles thousands of conversations, a human cannot sit inside every one. The practical question is which decisions need a human in the loop and which need automated oversight watching the rest.
Related: AI Agent Supervision, Agentic AI, Agent Policy & Guideline Enforcement
See also: What Is AI Agent Governance?