Talk to Your Data: RAG Server for Real-Time Intelligence in Microsoft Fabric

What if your real-time data could respond to questions in plain language? With the RAG Server for Microsoft Fabric, that’s now within reach.

In this practical session, Frank Geisler and Hanna Schwab show how to integrate natural language querying into Real-Time Intelligence workloads using Microsoft’s open-source RAG Server. Instead of writing complex KQL queries, users can interact with streaming data through simple, conversational prompts.

We’ll kick things off with an overview of how RAG and RAG Server work together, followed by a step-by-step setup: deploying the server, connecting it to a Fabric Eventstream, and demoing real-time queries over live data. The session emphasizes practical implementation, system architecture, and examples that can be applied in real-world scenarios.

If you’re working with Microsoft Fabric and exploring the use of large language models in analytics, this session provides a hands-on starting point.

Understand how RAG and the open-source RAG Server integrate with Microsoft Fabric Real-Time Intelligence.

Learn how to deploy and configure the RAG Server in a Fabric-based architecture.

Connect the RAG Server to Fabric Eventstreams for querying live, streaming data.

Execute natural language queries against real-time data without writing KQL.

Apply practical patterns and architecture concepts to real-world analytics scenarios.