From Vectors to Agents: Building AI-Powered Solutions with Microsoft SQL
Proposed session for SQLBits 2026TL; DR
Unlock the power of your own data by combining Microsoft SQL (SQL Server 2025, Azure SQL, or SQL Database in Microsoft Fabric) with cutting-edge AI techniques. In this workshop, you’ll learn about embeddings and vector search and why they matter for modern AI applications. We’ll guide you through generating embeddings using similarity search to retrieve the most relevant data.
We’ll implement the Retrieval-Augmented Generation (RAG) pattern by calling a Chat model, then explore its limitations. Evolve into Agentic RAG and create an orchestrator to intelligently choose between running semantic search or writing a SQL query to answer complex requests. Learn to secure data access so that only authorized information is available to AI agents using Row-Level Security
By the end, you’ll have built an end-to-end solution that combines SQL’s reliability with AI’s intelligence and be well prepared for SQL AI Developer certification exam.
Session Details
Unlock the power of your own data by combining Microsoft SQL (SQL Server 2025, Azure SQL, or SQL Database in Microsoft Fabric) with cutting-edge AI techniques. In this workshop, you’ll learn about embeddings and vector search and why they matter for modern AI applications. We’ll guide you through generating embeddings using similarity search to retrieve the most relevant data.
We’ll implement the Retrieval-Augmented Generation (RAG) pattern by calling a Chat model, then explore its limitations. Evolve into Agentic RAG and create an orchestrator to intelligently choose between running semantic search or writing a SQL query to answer complex requests. Learn to secure data access so that only authorized information is available to AI agents using Row-Level Security
By the end, you’ll have built an end-to-end solution that combines SQL’s reliability with AI’s intelligence and be well prepared for SQL AI Developer certification exam.
We’ll implement the Retrieval-Augmented Generation (RAG) pattern by calling a Chat model, then explore its limitations. Evolve into Agentic RAG and create an orchestrator to intelligently choose between running semantic search or writing a SQL query to answer complex requests. Learn to secure data access so that only authorized information is available to AI agents using Row-Level Security
By the end, you’ll have built an end-to-end solution that combines SQL’s reliability with AI’s intelligence and be well prepared for SQL AI Developer certification exam.
3 things you'll get out of this session
• Learn embeddings, vector search, and similarity search for AI applications
• Implement RAG patterns and evolve into Agentic RAG with orchestration
• Secure data access for AI agents using Row-Level Security
Speakers
Nikola Zagorac's other proposed sessions for 2026
Building event-driven apps usinng real-time change event streams in SQL - 2026
Dive into the engines that power Fabric Real-Time Intelligence - 2026
Zlatko Knezevic
Zlatko Knezevic's other proposed sessions for 2026
New in modern Developer Experiences in SQL - 2026
Performance monitoring for SQL, from ground to cloud - 2026