Human-in-the-loop (HITL) is a design pattern where a person reviews, approves, or can override an AI system's decisions at defined points. Rather than full autonomy, the agent handles the routine work and routes judgment calls, exceptions, or high-stakes actions to a human — preserving accountability and control.
Also known as: HITL, human oversight
Effective HITL puts review at the moments that matter: before an irreversible or high-risk action, on low-confidence outputs, and on anything with legal, financial, or safety significance — while letting routine steps run automatically.
We map escalation paths and approval checkpoints into every workflow, so agents accelerate work without removing human accountability.
The policies, controls, and processes that keep AI systems safe, compliant, transparent, and accountable across their lifecycle.
Read definition →Agentic AIA software component that pursues a goal by perceiving context, reasoning about what to do, and acting through tools and APIs — then iterating on the result.
Read definition →Agentic AIAI systems that take action — applying logic, making decisions, and running multi-step workflows end to end across enterprise systems, with human oversight.
Read definition →GovernanceIBM's toolkit to govern AI across its lifecycle — model risk, lineage, drift and bias monitoring, and policy controls with audit trails.
Read definition →