Human Oversight · Data artifact
Escalation Protocol
Data artifactHuman OversightExperience & Human OversightVariation point (abstract)arc:EscalationProtocol
An abstract specification of what the system does when an agent approaches or crosses a decision boundary, naming notified roles, decision owner, response-time objective and fallback behaviour.
Responsibility. Defines how a boundary crossing is handed to humans.
Also known as: Escalation playbook
Variants
| Variant | When to choose |
|---|---|
| Approval Escalation Protocol | Choose when the decision carries significant consequences but does not require immediate action. |
| Emergency Escalation Protocol | Choose when agents detect high-severity situations (security incidents, system failures affecting critical operations) requiring urgent human attention. |
| Notification Escalation Protocol | Choose when stakeholders need awareness of a boundary event but the action may proceed (e.g., internal transfers above $10,000 with enhanced logging). |
Relationships
configures structural
Design guidance
- MUST document clear ownership: which roles are notified and who holds decision authority.
- SHOULD set response-time service-level objectives for human responses.
- MUST specify fallback behaviour (decline, queue for retry, escalate further) when humans do not respond in time.
Classification
- Patterns
- Human-over-the-loop escalation
- Quality attributes
- Transparency and accountability (NIST AI RMF: accountable and transparent)Performance efficiency (ISO/IEC 25010)
- Risks mitigated
- Agents left in indefinite or ambiguous waiting statesUnclear decision ownership
Sources
- Ch10.5: T. Nguyen, "Human-over-the-Loop," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.5. ISBN: 9798244538229.