Ontology
Last updated June 14, 2026
What is Ontology in simple terms?
In simple terms, an ontology is a clearly written map of how the things in a subject relate — what counts as what, and how they connect. Like a family tree that spells out who's a parent, sibling, or cousin.
What is Ontology?
An ontology, in computing and AI, is a formal, machine-readable description of the concepts in a subject area and the relationships between them, defining the categories, properties, and rules that let software interpret information consistently rather than treating it as plain text.
The word "ontology" comes from philosophy, where it means the study of what exists. In computing and AI it's borrowed for a more practical job: writing down, in a way software can use, what the things in a particular subject *are* and how they *relate*. An ontology defines the categories in a domain (a "car" is a kind of "vehicle"), the properties things can have (a car has a manufacturer, a number of wheels), and the relationships between them (a manufacturer *produces* a car; a car *requires* a driver). The point is to make meaning explicit. To a computer, the word "Apple" is just letters; an ontology can record that this Apple is a *company*, that it *makes* products, and that a product is a *different kind of thing* from the company — so the software can reason about it rather than just match strings.
This explicitness is what separates an ontology from an ordinary list or spreadsheet. A spreadsheet of fruits and companies sits there as data; an ontology adds the *structure of meaning* — the hierarchy of categories, the allowed relationships, the rules ("a person can manage a company; a fruit cannot"). With that structure in place, software can draw conclusions that were never written down directly: if it knows a robin is a bird and birds can fly, it can infer a robin can fly without anyone stating it. Ontologies are typically built in standardized formats so that different systems can share and combine them, which is part of how knowledge can be pooled across organizations rather than locked in one company's private format.
Ontologies are closely tied to knowledge graphs and knowledge bases. A useful way to see the relationship: an ontology is the *schema* — the agreed vocabulary and rules of what kinds of things and links are allowed — while a knowledge graph is that vocabulary filled in with actual facts and entities. The honest caveats are about effort and agreement. A good ontology takes real expertise to design, must be kept up to date as a field changes, and only works if people agree on its definitions — two teams that model the same domain differently can end up with incompatible ontologies. It's a powerful way to give machines structured meaning, but it's a deliberate, maintained piece of engineering, not something that builds itself.
Real-world example of Ontology
A hospital network wants its software to connect information that currently lives in separate systems — patient records, a drug database, and a list of conditions. The trouble is that nothing agrees on meaning: one system calls it "paracetamol," another "acetaminophen," and the records don't know these are the same medicine. The team builds a medical ontology that states it plainly: paracetamol and acetaminophen are names for *one* drug; that drug *treats* certain symptoms; those symptoms *belong to* certain conditions; some drugs *interact dangerously* with others. Now the software can do real reasoning — flag that a newly prescribed drug clashes with one a patient already takes, even though the two were recorded under different names in different systems. The ontology gave the machine a shared, explicit map of what these medical things are and how they relate.
Related terms
Frequently asked questions about Ontology
What is the difference between an ontology and a knowledge graph?
They work together but play different roles. An ontology is the *blueprint*: it defines the kinds of things that can exist in a domain, their properties, and the relationships allowed between them — the rules and vocabulary. A knowledge graph is the *filled-in version*: actual entities and facts connected according to that blueprint (this specific patient, this specific drug, this specific link between them). In short, the ontology says "people can work for companies"; the knowledge graph says "Maria works for Acme." One defines the structure of meaning; the other populates it with real data. **2. Mechanism — How does an ontology work?**
How does an ontology work?
It captures a domain as a formal set of definitions a machine can read: categories arranged in hierarchies (a dog is a kind of mammal is a kind of animal), the properties each category can have, the permitted relationships between them, and logical rules. Written in standardized formats, this lets software interpret data by meaning rather than as plain text — and crucially lets it *infer* facts that weren't stated directly, by following the defined hierarchies and rules. Because the structure is explicit and shared, different systems can combine their information instead of each guessing at what the other's data means. **3. Application — What is an ontology used for?**
What is an ontology used for?
Ontologies are used to give machines a shared, structured understanding of a subject so they can integrate data, reason over it, and exchange it between systems. They underpin knowledge graphs (including those behind search engines and digital assistants), help merge information across departments or organizations that store it differently, and are widely used in fields with complex, precise vocabularies — medicine, biology, law, and engineering. Anywhere meaning must be unambiguous and machine-usable rather than left to plain text, an ontology supplies the formal backbone.