AI models can generate syntactically correct SQL but often fail to return meaningful results because they lack business context. Kris Jenkins, Lead Developer Advocate at Snowflake, presents a solution using semantic models and the OSI (Open Semantic Interface) standard. These tools provide AI agents with a structured understanding of what database tables represent, how entities relate, and what business rules apply — bridging the gap between SQL syntax knowledge and domain-aware query generation. The talk covers how to implement semantic models declaratively, why the OSI standard matters, and how to build a reliable AI-powered database analyst.
Table of contents
Context of the TalkSemantic Models and the OSI Standard for AI AgentsThe TalkAbout the Speaker: Kris Jenkins, Lead Developer Advocate at SnowflakeSort: