Beyond Vector Databases: Structured Retrieval and Graph-Native AI Systems

This Tokyo AI workshop looks past the now-standard RAG stack of embeddings and vector databases, which can become hard to maintain when data is highly structured, relationship-heavy, and constantly changing. Across two technical sessions, Adam Gibson explores retrieval architectures built on relational databases, graph structures, ontologies, keyword search, and tool-mediated retrieval to ground AI agents without depending on embeddings alone. Part 1 presents a production architecture that grounds agents entirely through structured retrieval, using an AI-powered virtual tabletop RPG platform as a case study for relational grounding, entity resolution, prioritized context assembly, and MCP-style retrieval. Part 2 turns to graph-native systems, showing how ontologies and knowledge graphs enable more precise retrieval, explainable reasoning paths, and dynamic exploration by agents. The sessions target engineers, researchers, and technical leaders building production AI systems, enterprise knowledge platforms, agentic workflows, or retrieval infrastructure. The evening opens at 18:00, runs two talks until 20:00, and closes with an hour of networking.