Automatically map your business logic, metrics, and definitions across all your systems
Continuously resolve mismatches, anomalies, and schema drift without manual intervention
Deliver structured, contextualized intelligence directly to copilots, dashboards, and decision workflows
A data ontology is a formal model of the things a business cares about, such as customers, accounts and products, and how they relate to each other. A glossary defines terms one at a time. An ontology also defines the relationships, so people and AI can reason across systems that describe the same things differently.
The term comes from philosophy and computer science. It's back in use because AI agents need it.
What an ontology contains
- Concepts. The kinds of things that exist, such as customer, account, product and GL account.
- Relationships. How concepts connect: a customer holds an account, and an account is an instance of a product.
- Attributes. What describes each concept, such as an account's open date or balance.
- Rules. Constraints that must hold, such as every account belonging to exactly one legal entity.
- Mappings. Where each concept lives in each system.
How it differs from related terms
- Glossary. Defines terms in plain language. No relationships.
- Taxonomy. Arranges terms in a hierarchy, such as a chart of accounts. Relationships are only parent and child.
- Data model. Describes how one system stores data. An ontology describes the business, across systems.
- Knowledge graph. Stores actual facts as connected entities. An ontology usually provides its structure.
- Semantic layer. Defines metrics on warehouse tables for reporting. It can be built on an ontology but rarely captures all the relationships.
Why finance needs one
"Deposits" means one thing in the core, another in the GL and a third in the FP&A model. Each system is right for its purpose. An ontology records how those versions relate to the same concept, so a question about deposits by relationship can travel from customer to account to product to balance without guessing.
How Preql helps
Preql's agents find how each of your systems represents the same business concept and turn those representations into shared business logic. Business owners approve the mappings, and every change is versioned with an audit trail and lineage.
