Agentic AI Reference Architecture

Human Oversight · Software component

Human Validation Checkpoint

Software componentHuman OversightExperience & Human OversightVariation point (abstract)arc:HumanValidationCheckpoint

An abstract oversight checkpoint at which humans validate agent actions, either before execution (approval-before) or after execution on completed actions (approval-after).

Responsibility. Places human validation relative to action execution.

Also known as: Decision approval model, Approval pattern

is specialized byguardsis specialized byApproval Gateway: is specialized byApproval GatewayProactive Agent: guardsProactive AgentPost-Action Review Sampler: is specialized byPost-Action Review Sampler
Direct neighbourhood (hover for relationship types)

Variants

VariantWhen to choose
Approval GatewayChoose for high-consequence irreversible decisions (financial transactions, contractual commitments, changes affecting many users) where errors must be prevented before impact, accepting reduced throughput.
Post-Action Review SamplerChoose for lower-stakes reversible actions where operational velocity matters; requires the ability to reverse completed actions when reviews reveal errors.

Relationships

guards control

Design guidance

Classification

Patterns
Approval-beforeApproval-after
Quality attributes
Transparency and accountability (NIST AI RMF: accountable and transparent)Safety (ISO/IEC 25010 | NIST AI RMF: safe)Performance efficiency (ISO/IEC 25010)

Sources

  1. Ch10.2: T. Nguyen, "Proactive Agents," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.2. ISBN: 9798244538229.