SQL Server 2025 introduces native support for vector storage and AI-driven query capabilities, fundamentally changing how generative AI solutions can be built within the database engine. Rather than exporting data to external platforms or introducing complex data movement pipelines, organizations can now design and operationalize AI workloads directly where their data already resides.
In this session, you will learn how to leverage SQL Server 2025’s built-in vector architecture to store embeddings, perform similarity searches, and power retrieval-augmented generation (RAG) scenarios using your enterprise data. We will walk through the new T-SQL AI syntax that allows you to integrate large language models into standard query workflows, enabling tasks such as semantic search, summarization, and contextual question answering without leaving the SQL environment.
You will see how to configure and connect to the AI model of your choice, including how to define model endpoints, control where models are hosted, and manage inference execution within SQL Server. The session focuses on practical implementation, showing how to design production-ready patterns that combine relational data with vector-based intelligence while maintaining governance, security, and performance.
By the end of this session, you will understand how to build, deploy, and scale generative AI solutions natively in SQL Server 2025, eliminating unnecessary architecture complexity and bringing AI directly to your data platform.