The missing layer in your AI stack

Most AI applications find information by similarity, which works well until a question depends on how things relate. Brian O'Keefe, Solutions Engineering Manager for the Neo4j Startup Program, will introduce graph databases to a general audience and explain how they store data as connected entities and relationships. He will show how graphs sit at the center of three of the hottest topics in AI: agentic memory, explainability, and GraphRAG.

Organizations across Rochester's healthcare, education, business, and research communities hold data where connections matter: patients and care pathways, courses and prerequisites, suppliers and products, studies and citations. When an AI system cannot see those connections, it returns plausible answers that it cannot explain. This session will make the case that connected data gives AI agents durable memory, gives users traceable answers, and gives retrieval the context that vector search alone lacks.

You will leave with a clear mental model of how graphs fit into an AI architecture and an understanding of where they improve accuracy, explainability, and reasoning. You will also take away a short list of use cases to evaluate in your own field. No prior experience with graph databases or AI development will be required.