What is a process ontology?

A process ontology is a formal description of the concepts in a process domain and their relationships: which object types exist, which events occur on them, and how the two connect. It records that a purchase order, a delivery, and an invoice are distinct objects with defined links – not merely column names in a table. This turns raw data into a model that machines can interpret as well.

Why does this matter?

AI agents can only work as precisely as the meaning of the data is recorded. Without an ontology, every model has to guess the semantics; with one, it can answer questions along real relationships and avoid false conclusions from wrongly linked objects.

What does this look like in practice?

The ontology records that an invoice relates to one or more orders, that "cancelled" is a state of the order and not of the delivery, and that "approval" means the same event in purchasing as in sales. Without those commitments, every report counts something slightly different.

Not to be confused with

Not to be confused with a data model, which defines tables and fields; the ontology defines the meaning behind them. Nor with a glossary, which explains terms for people — an ontology is machine-readable and can be reasoned over.