New! Relate by Neo4j | An embedded database for AI agents Learn More

NODES 26 — November 12, 2026

Building a B2B Context Graph: Modeling Relationships and Outcomes for AI Agents

Session track: AI Engineering

Session time:

Session description:

Everyone says relationships close deals. Almost nobody can prove it, because the evidence, who knows whom, which events they met at, and which interactions influenced revenue, lives scattered across CRMs and event platforms in shapes a relational schema struggles to express. This session is about what happens when you model that evidence as a graph. Ben Roodman (CEO) and Anthony Perritano, PhD (CTO) of heyBTW will share how they built a B2B Context Graph on Neo4j Aura that connects event platforms and CRMs into a single relationship layer. They will walk through the five-coordinate graph data model (Events, Timeline, Relationships, Outcome, Attribution), the "one canonical resolver per coordinate" pattern that keeps every surface, including AI agents, reading the same truth, and the anti-patterns they learned to avoid the hard way. You will see the Postgres-to-Neo4j sync architecture, real Cypher queries behind multi-touch attribution traversals, and how the graph is exposed to AI agents through an MCP server with read tools in production today. You will leave with a reusable blueprint for modeling business relationships as a graph and making that context available to agentic systems, whether your domain is GTM or anything else where relationships drive outcomes.

Speaker

photo of Benjamin Roodman

Benjamin Roodman

CEO & Co-Founder, heyBTW

Ben Roodman is CEO and Co-Founder of heyBTW, the B2B Context Graph that proves which relationships close deals. A 16-year MarTech veteran, he previously led an 11-person partnership team at AppsFlyer through 100x growth ($4M to $400M ARR) and has built partner ecosystems of 10,000+ integrations across three platforms. After running more than 100 B2B events, he started heyBTW to solve the problem he lived: nobody could prove which relationships and events actually drove pipeline.