In mathematics and computer science, graph theory studies mathematical structures that model pairwise relations between objects called graphs. Thus, in computer science, a graph database is a NoSQL Database that uses graph structures to store and query connected data.

The objects within the graph database are:

• Node — Represents entities, such as a person, product, or place.
• Edge — Connects two nodes and represents their relationships.
• Property — A key-value pair that adds details to a node or an edge.

Therefore, graph databases store and maintain relationships among data over time, rather than inferring them when needed, as in relational databases.

A knowledge graph captures what is connected, and vector databases retrieve what appears similar. Furthermore, a context graph extends the knowledge graph by adding time and decision lineage. While a knowledge graph knows what a customer is, a context graph captures the specific reasoning behind a decision made for that customer at a specific moment. In other words, context graphs record the dynamic procedural logic and event traces, which are essential for Agentic AI to avoid “context rot” and hallucinations.