Beyond the Experience Graph: The Data Graph Meets AI

Most of us know GraphQL as the "Experience Graph", shaped around screens and clients. But GraphQL can be used elsewhere in your architecture: as a Data Graph, a strongly-typed abstraction directly over your data layer, giving every consumer - services and AI agents alike - a single place to access all your data.

In this talk, I'll show how modelling your databases behind a GraphQL schema gives you a self-describing, introspectable contract and why that turns out to be a remarkably good surface for AI. Through demos, we'll see AI agents introspect a Data Graph to explore data, assist developers in writing and refining queries, and even propose schema improvements by spotting gaps and ambiguities humans miss.

You'll leave with a clear mental model of when a Data Graph makes sense, how it differs from the experience graph you already know, and practical patterns for making your graphs AI-ready.