COMPEL Glossary / agent-registry
Agent Registry
An agent registry is a centralized catalog that tracks all deployed AI agents across an organization, documenting their capabilities, permissions, owners, operational status, tool access, and governance compliance.
What this means in practice
Similar to how a model registry tracks AI models, the agent registry provides the single source of truth for understanding what autonomous systems are operating within the enterprise. Without a registry, organizations risk losing track of agents that may be taking actions, consuming resources, or making decisions without appropriate oversight. The agent registry is a foundational governance control for agentic AI discussed in COMPEL Module 3.4, Article 11, enabling the governance board to maintain visibility and control over the growing population of autonomous AI capabilities deployed across business units.
Why it matters
Without a centralized registry, organizations lose track of autonomous AI systems operating across business units, creating invisible risk exposure. Each unregistered agent represents a potential compliance violation, security vulnerability, or operational disruption that the governance board cannot manage because it does not know the agent exists. The agent registry is the foundational visibility control that makes all other agentic AI governance possible.
How COMPEL uses it
The agent registry is established during the Produce stage as a foundational governance control for agentic AI within the Governance pillar. During Calibrate, existing agents are discovered and cataloged as part of the technology landscape assessment. The registry tracks capabilities, permissions, owners, operational status, and governance compliance, enabling the Evaluate stage to audit agent behavior and the governance board to maintain oversight of the growing agent population.
Related Terms
Other glossary terms mentioned in this entry's definition and context.