Most GraphRAG systems retrieve context for an LLM. John Thomas and the GraphLogic team have been exploring what happens when the reasoning itself lives in the graph.

On Thursday, July 2, John will walk through the reasoning platform GraphLogic has built on Memgraph — where arguments, evidence, risks, assumptions, and decisions are stored as graph structures instead of disappearing into prompts and chat history.

The session tackles a problem most organizations already have: important decisions are scattered across meetings, tickets, architecture reviews, and disconnected systems. The data exists, but the reasoning behind each decision usually doesn't. GraphLogic uses Memgraph as the operational graph behind that reasoning, so every decision can be traced back through its evidence, competing viewpoints, dependencies, and source materials. John will also demonstrate how Memgraph Zero acts as a federated query layer across traditional databases, letting the graph reference enterprise evidence without centralizing the data.

For technology leaders: a model for turning scattered decisions into auditable, long-lived institutional knowledge — reasoning that survives past the meeting it happened in.

For engineers and architects: how that reasoning is actually modeled and queried — coordinating graph traversal, vector search, analytics, and LLM calls in one operational layer.

In this session, we'll cover:

How GraphLogic pulls external enterprise data into a common context graph for structured reasoning

Extending graph queries with temporal context, change events, and document-aware features

How the platform captures organizational learning over time through the LEAP™ cycle (Learn / Envision / Act / Perfect)

This session is for graph engineers, architects, AI practitioners, and technology leaders interested in how graph systems move beyond retrieval into long-lived organizational reasoning.

How GraphLogic built a traceable reasoning layer on Memgraph