Ontology

A formal representation of entities and their relationships within a domain, used to encode structured knowledge that can be shared with and reasoned over by AI systems. In the context of agent engineering, ontologies serve as the conceptual backbone of Knowledge Graphs.

Definition and Role

Frank Coyle (UC Berkeley) defines ontologies functionally: "They're basically a representation of entities and their relationships to other entities" — establishing them as the structural layer that gives Knowledge Graphs their semantic meaning. ↗

The Case for Ontologies in Agentic Systems

Coyle argues that as AI agents are deployed in real-world domains, they need more than raw data or unstructured text — they need a grounded conceptualization of the world they operate in. His central advocacy position is that practitioners should actively supply agents with domain models: "We want to give them our conceptualization of the universe, our universe, our domains." ↗

This framing positions ontologies not as a passive data artifact but as an intentional design decision by agent engineers — a way of constraining and grounding agent reasoning within a well-defined semantic space. The fit between agentic systems and ontologies is the core thesis Coyle advances: "I want to talk about how agents and ontologies will — big word — fit together." ↗

Points of Disagreement

No dissenting positions on ontologies are recorded in the current graph. The node is represented exclusively through Coyle's advocacy.

See Also

• Knowledge Graph — the structure that ontologies define