GraphQL meets the agent era: Matt Debergalis on APIs, MCP, and enterprise AI

Apollo GraphQL's schema and type system offer a structured, typed layer over API surfaces, which is crucial for AI agents to understand the relationships and meanings of data, rather than just endpoints. At HumanX 2026, Apollo co-founder Matt DeBergalis highlighted the synergy between GraphQL and Anthropic's Model Context Protocol, emphasizing that while MCP handles the connection mechanics, GraphQL describes the data, enabling agents to construct meaningful queries. This approach allows enterprises to modernize without rewriting legacy systems, by making them accessible to AI through a typed schema layer. As enterprises adopt AI, the focus has shifted not only to technology but also to how people work and collaborate, reshaping workflows and organizational structures. The ability to introspect and query data within systems using tools like Claude Code is seen as unlocking hidden value, with non-developers increasingly building AI agents, showcasing a shift in skill requirements from traditional programming to systems design and problem-solving. This evolution is also impacting procurement, as the ease of building customized solutions is changing the logic from buying to building, marking a transformative period in enterprise technology strategy.