Knowledge graph
A knowledge graph is a way of storing knowledge as a network: entities (people, machines, clients, decisions) are nodes, and their relationships are the connections between them. Facts keep their context — who decided what, which machine belongs to which client — and can be traversed and queried like a map.
What is knowledge graph?
Prose buries relationships; a graph makes them first-class. "Supplier X was dropped in 2009 after quality issues on the P-40 line" becomes connected nodes — supplier, decision, product line, date, reason — each reachable from the others. Ask about the P-40 line and the supplier history is one hop away, even though no document ever put them in the same paragraph.
Not to be confused with a mind map or an org chart. Those are drawings made for one purpose. A knowledge graph is a data structure: machine-queryable, extensible, and able to answer questions its builders never anticipated.
Why it matters for SME continuity
An expert's knowledge is relational by nature — this client, that machine, that incident, this rule of thumb, all interlinked. Flat documentation cuts those links; readers must reconstruct them in their heads, which is exactly the skill that leaves with the expert. A graph preserves the links themselves, so a successor inherits the connections and not just the fragments.
Knowledge graph inside MentX
MentX is a personality-aware knowledge base built on a knowledge graph. Everything captured from meetings and the guided intake lands as connected, sourced facts. Two properties are added on top: time (see temporal knowledge graph) and provenance — every node knows when and where it came from.