Glossary

What is a product knowledge graph?

A product knowledge graph represents products and related entities through explicitly described relationships. It can connect items with brands, categories, attributes, compatible parts, documents and other information. The graph structure makes those connections available for queries and applications, but its usefulness depends on consistent identities, relationship meanings and reliable source data.

Beyond a category tree

A taxonomy places products in categories. A graph can represent several kinds of relationship at the same time, such as a part fitting an assembly, a document describing a model or a product replacing a discontinued item. These relationships need clear definitions and identifiers.

For example, a distributor could connect a pump with its compatible seal kits and the installation document that explains the fit. A buyer-facing application could query those relationships instead of attempting to infer compatibility from similar product names.

A graph is a model, not proof

RDF is one standards-based way to represent linked information using subject, predicate and object statements. Other graph technologies use different representations. A product knowledge graph does not have to be a particular vendor platform or a specific database implementation.

Adding a relationship does not make it correct. Record where important links came from, whether they are verified and what context limits them. Compatibility may depend on model revision, size or operating conditions. A graph can support search, retrieval and AI applications, but it still requires the same care with evidence and change control as other product-data structures.

Related terms

Sources and further reading

Will AI Recommend Your Products?

Explore the role of product data in AI discovery and recommendations.