Graph Database Boston

This hands-on workshop shows you how to build smarter AI agents: ones that don’t lose context. You’ll use Neo4j’s Model Context Protocol (MCP) to give your agents persistent, graph-based memory, allowing for richer, multi-turn interactions and intelligent behavior grounded in your data.

You’ll walk through the full agent-building process: from setting up a knowledge graph, to integrating GraphRAG for multi-hop reasoning, and deploying an agent with contextual memory that evolves over time. By the end of the session, you’ll have a working, production-ready AI agent you can extend for your own use cases.

How can we use graphs and AI to bring transparency to politics? In this live session, we’ll dive into Valkompass.ai - an open-source project that uses Neo4j and Gemini to connect and analyze political data from Sweden’s parties. From election manifestos to parliamentary votes, Valkompass.ai turns complex, unstructured information into a knowledge graph you can query and converse with.
We’ll explore how this system models political relationships, applies AI reasoning over connected data, and helps citizens and journalists navigate policies with context and clarity.

Guest: Emil Wåreus
Valkompass: https://www.valkompass.ai/
Github: https://github.com/valkompass-ai/valkompass-ai

Ever wanted to build an agent that can understand your data and answer complex questions for you? In this workshop, you’ll roll up your sleeves and create your own agent in Neo4j Aura, complete with a UI and a real use case around legal documents and KYC (Know Your Customers) checks.