Human-in-the-loop Gate

Human-in-the-loop Gate A sequence diagram generated by Archify. propose action routine: auto under guardrails high-consequence: human required approve within time box execute action reject or timeout: drop log approver, context, decision Classify by consequence Human decides in a time box Execute or drop, then log Agent · proposes action · Sequence participant Agent proposes action Policy check · human required? · Sequence participant Policy check human required? Reviewer · named human · Sequence participant Reviewer named human Action target · payment, denial · Sequence participant Action target payment, denial Evidence log · approver, decision · Sequence participant Evidence log approver, decision Legend request return security async trace default message

Oversight by design

  • • Classify actions by consequence, not by habit
  • • Gate only the irreversible or high-impact class behind a human
  • • Keep the routine class autonomous under guardrails

A decision in a time box

  • • The agent pauses and the action is blocked until a named reviewer decides
  • • The reviewer gets enough context to decide
  • • A timeout drops the action rather than letting it through

Recorded as evidence

  • • The approver, the context and the decision are logged
  • • This realises the EU AI Act Art. 14 human-oversight duty at the point of action
  • • The approval is auditable