A formal representation of knowledge as a set of concepts, entities, and their relationships.
In computer science and AI, an ontology is a formal, explicit specification of a shared conceptualization. It defines the types of entities that exist in a domain, the properties those entities can have, and the relationships between them. Unlike a simple database schema, an ontology captures the meaning and logic of a domain — enabling machines to reason about it, infer new facts, and understand context.
With the rise of agentic AI, ontologies have become mission-critical infrastructure. When an AI agent operates autonomously — booking travel, analyzing contracts, or managing inventory — it must understand what business concepts mean and how they relate. Without a well-defined ontology, agents make logical errors that compound over time. The industry has recognized that ontology engineering is now a core AI engineering discipline.
Ontologies are typically expressed in formal languages like OWL (Web Ontology Language) or RDF (Resource Description Framework). They define classes (types of things), individuals (specific instances), properties (attributes and relationships), and axioms (logical rules). Ontology reasoners can then infer new facts — for example, if an ontology defines that 'all mammals are warm-blooded' and 'dogs are mammals,' a reasoner can infer that 'dogs are warm-blooded' without that fact being explicitly stated.
A healthcare organization builds an ontology that defines 'Patient,' 'Diagnosis,' 'Medication,' and 'Treatment' as classes, with properties like 'hasCondition,' 'prescribedBy,' and 'contraindicatedWith.' An AI agent using this ontology can reason that a patient with a specific condition should not receive a particular medication — even if that rule was never explicitly programmed.