Multi-Agent Shared Graph Memory: Building Collective Knowledge for Agents

Session Track: Graph Memory & Agents

Session Time: 11:00 am - 11:30 am April 15

Session description

In the age of autonomous systems, AI agents no longer act alone-they collaborate, negotiate, and evolve shared understanding. In this session, Vaibhava Lakshmi Ravideshik will explore how multiple agents can reason over a common graph-based memory, enabling persistent, explainable, and cooperative intelligence.

You will learn how to design a shared knowledge graph architecture that agents can update and query in real time-complete with conflict resolution, versioning, and provenance tracking. The talk will feature live code snippets and Cypher examples demonstrating how Neo4j can serve as the backbone of collective agent memory.

By the end, you’ll understand how to:
– Represent agent beliefs and updates in a shared graph
– Use graph queries for decision-making and context recall
– Manage evolving multi-agent knowledge safely and transparently

Whether you’re building RAG systems, AI assistants, or simulation environments, this session will show how to give agents the gift of shared memory-and make them smarter together.