Skip to main content
COMPEL Body of Knowledge

COMPEL Glossary / GL-28

Agent Autonomy Classification

A formal classification of every AI agent in scope according to its autonomy level — from level 0 (no autonomy, human executes) through level 4 (full autonomy, agent executes without human involvement) — with corresponding governance requirements, approval boundaries, and monitoring obligations assigned at each level.

What this means in practice

It prevents high-autonomy agents from operating without appropriate oversight.

Context in the COMPEL framework

Produced in the Model stage for any AI system involving autonomous or semi-autonomous agents. Classification at this stage determines the agent governance controls required in the Produce stage.

Where you see this

Agent Autonomy Classification is most commonly referenced when teams work across the Model , Produce and Evaluate stages — especially within the Agent Governance layer . It appears in governance artifacts, assessment instruments, and delivery playbooks wherever COMPEL is operationalized.

Related COMPEL stages

Related domains

Synonyms

autonomy level register , agent risk classification , agentic autonomy tiers

Related Terms Network

This diagram shows how Agent Autonomy Classification connects to other concepts in the COMPEL body of knowledge. Each linked node is a term you can explore in turn.

Related terms for Agent Autonomy Classification Network graph showing 3 related glossary terms connected to Agent Autonomy Classification. Agent Autonomy… AI System … Human Vali… Control Re…
3 related concepts connected to Agent Autonomy Classification.
Accessible list of related terms
  • AI System Classification Register — A formal register that classifies every AI system in scope according to risk tier, autonomy level, data sensitivity, regulatory applicability, and criticality — producing a system-level risk profile that determines which governance controls, review processes, and compliance requirements apply.
  • Human Validation Rules — A formal specification of every decision point in an AI system's operation where human review, approval, or override is required — including the trigger conditions, review timeline, escalation path, and documentation requirements for each rule.
  • Control Requirements Matrix — A comprehensive mapping of every governance control required for each AI system — specifying the control type (preventive, detective, corrective), the risk or policy it addresses, the evidence required to prove effectiveness, the owner, and the testing frequency.

See also

  • AI System Classification Register — A formal register that classifies every AI system in scope according to risk tier, autonomy level, data sensitivity, regulatory applicability, and criticality — producing a system-level risk profile that determines which governance controls, review processes, and compliance requirements apply.
  • Human Validation Rules — A formal specification of every decision point in an AI system's operation where human review, approval, or override is required — including the trigger conditions, review timeline, escalation path, and documentation requirements for each rule.
  • Control Requirements Matrix — A comprehensive mapping of every governance control required for each AI system — specifying the control type (preventive, detective, corrective), the risk or policy it addresses, the evidence required to prove effectiveness, the owner, and the testing frequency.

Related articles in the Body of Knowledge

Related Terms

Other glossary terms mentioned in this entry's definition and context.