22-25 April 2026 
Conversational Experiences in Microsoft Fabric with Data Agents
Proposed session for SQLBits 2026TL; DR
Learn how to build and publish Fabric Data Agents in Microsoft Fabric. This session covers core AI concepts, a live demo, and best practices for data prep and schemas to create effective, customized AI agents.
Session Details
LLMs and Agentic AI have already made their way into Microsoft Data Platform and one of the newest implementations are Fabric Data Agents. They enable you to create a highly customized AI Agent within Fabric that you can feed with custom context, selected data sources and pretrain to fit your needs. You can then publish the AI Agent to make it available for your colleagues to interact with.
During this session, you will learn the core Fabric & AI concepts necessary to work with Fabric data agents, know how to create a data agent during a live demo as well as be familiar with the best practices which will make your agent even better, such as the right data preparation and AI data schema.
As a prerequisite before this session, it would be great if you have a working experience with Power BI / Fabric / SQL or familiarity with core data modelling concepts.
During this session, you will learn the core Fabric & AI concepts necessary to work with Fabric data agents, know how to create a data agent during a live demo as well as be familiar with the best practices which will make your agent even better, such as the right data preparation and AI data schema.
As a prerequisite before this session, it would be great if you have a working experience with Power BI / Fabric / SQL or familiarity with core data modelling concepts.
3 things you'll get out of this session
After this live demo-driven session you will:
- know the core concepts behind LLMs and Agentic AI
- be able to implement a Fabric Data Agent in your own environment
- be familiar the latest art of possible with Fabric Data Agents