Your multi-agent procurement workflow can draft purchase orders and call a create-PO tool. Finance requires that any PO above $10,000 pause for human approval before the tool runs, with the ability to override or reject. How should you implement this control?
Select an answer to reveal the explanation.
Short Explanation
Answer A. AI-500 explicitly covers human-in-the-loop processes: approval workflows, overrides, and edge-case handling. Gate the create-PO tool when amount exceeds the threshold, persist workflow state during the wait, and resume only on approve—or exit cleanly on reject/override. Temperature tweaks do not create approvals. After-the-fact email is not a preventive control. Removing tools entirely abandons automation below the threshold.
Full Explanation
Correct answer: A. Production multi-agent systems combine autonomy with HITL for high-impact actions. Implement an approval node (Agent Framework / workflow engine) that intercepts tool calls above policy limits, notifies approvers, stores pending state, and only then allows the tool to execute. Support explicit reject and human override paths for exceptions.
B is incorrect: sampling temperature is unrelated to authorization or approval semantics.
C is incorrect: post-creation notification is detective, not preventive, and fails the “before the tool runs” requirement.
D is incorrect: permanently removing tools over-constrains low-risk automation and is not a selective HITL design.
Pair HITL with audit logs, correlation IDs, and clear agent personas that know when they must escalate.