Suspend and Resume
Suspend and resume is a governance capability in which an AI agent pauses mid-task when it encounters an action that requires human approval, waits for a policy decision, and then continues with full context once the decision resolves. The agent experiences the pause as a slow tool call, preserving its working state.
Why It Matters
Most agent approval mechanisms are binary: the user clicks "allow" or "deny" before the action executes. This works for simple flows but breaks in multi-step tasks. If an agent is ten steps into a complex workflow, a binary prompt forces the user to approve blindly or restart.
Suspend and resume decouples the approval decision from the agent's execution state. The agent pauses, the approval request is routed to the appropriate person (a manager, security reviewer, or app owner), and the agent resumes exactly where it stopped.
This is distinct from MCP elicitation (the protocol method for inline approval prompts) and from confirmation dialogs in chat UIs. Suspend and resume is organizational governance: the approval is evaluated against policy, not user preference.