Fabric Data Agent + Foundry Agent: Building a Multi-Source AI Assistant

In this video we'll take a look at a Microsoft Fabric Data Agent extended by integrating it into a Microsoft Foundry Agent to create a powerful multi-source AI assistant. First, we explore how a Fabric Data Agent can answer complex questions over a Fabric Ontology, including queries that traverse multiple entity relationships like Drivers, Trips, Loads, Customers, and Terminals. Then we take things further. Inside Microsoft Foundry, we'll see an agent that combines multiple knowledge sources: • The Fabric Ontology Data Agent for structured enterprise data • An Azure AI Search vector index containing company documentation This allows a single agent to answer questions from both structured data and unstructured documents. We demonstrate: • Querying operational data through the Fabric Ontology • Calculating load value for trips currently in progress • Answering company policy questions from documentation • How Foundry orchestrates multiple knowledge sources Finally, we discuss how this type of agent can be embedded into custom applications, enabling enterprise systems to support natural language queries over both data and documentation.