Context graphs
A context graph connects a company's entities and their relationships instead of storing knowledge as isolated documents.
Category
Industry concept
See all terms
What is a context graph?
A context graph represents a company's knowledge as entities, like customers, employees, products, and projects, connected by the relationships between them. A contract, a customer, and the deal it belongs to aren't three unrelated files; a context graph keeps that connection explicit, so an agent can follow it.
How does a context graph differ from a knowledge graph?
A knowledge graph is a general-purpose way to represent entities and relationships, and it powers very different use cases: a search engine's knowledge panel, fraud detection in banking, supply chain visibility. A context graph is a knowledge graph designed to providing AI agents with a company's context. It requires entities and relationships that map onto how the company actually operates, kept current through continuous updates and readable through the same permissions an agent already has.
How does a context graph work?
A context graph is constructed by pulling records out of connected sources, normalizing them into entities, and linking entities that reference each other, so a customer mentioned in a support ticket and the same customer's row in the CRM collapse into one connected record. As sources change, the graph updates, and duplicate entities that show up across systems get resolved into a single node. This is what a context repository holds underneath its file structure: not isolated files, but a graph of connected, current entities that agents and users can read through the same permissions.