What Is a Knowledge Graph (and What It Isn’t)

Knowledge graph definition

A knowledge graph is a way of modelling reality where:

• Entities are nodes

• Relationships are edges

• Meaning comes from how things connect — not just what they are

In a graph:

• Nodes exist to be connected

• Relationships are stored directly

• Traversal is native, not simulated

Nodes and edges explained simply

Nodes represent things that exist:

• People

• Accounts

• Systems

• Groups

• Roles

• Applications

Edges represent how those things relate:

• Membership

• Ownership

• Entitlement

• Trust

• Dependency

Edges have direction and meaning.

Risk rarely lives inside a node.
It emerges between nodes.

Traversal: where insight comes from

Graphs are designed to answer questions like:

• What paths exist between these identities?

• How many hops does privilege travel?

• Where does access fan out or converge?

This process is called traversal.

Traversal answers “how things connect” — not just “what they are”.

Familiar examples outside identity

Knowledge graphs are already used where relationships matter most:

• Fraud detection (money flows)

• Social networks (influence paths)

• Network routing (path optimisation)

• Recommendation engines (relationship inference)

Identity has the same characteristics — but far higher risk.

What a knowledge graph is not

A knowledge graph is not:

• A visualisation layer

• A reporting trick

• A replacement for source systems

• A magic AI engine

It is a structural choice about how reality is represented.

➡️ Next: How Gathid applies this model specifically to identity.