Summary
Unit 7 of 7 · Module 10: Build knowledge-enhanced AI agents with Foundry IQ
Read this lesson
Five lanterns
That's Module 10, so let's walk the pier at dawn and light one lantern for each idea. The first is RAG: an agent that looks things up in your own knowledge before it answers, instead of guessing from memory. The second is Foundry IQ, the shared knowledge layer: knowledge bases built once, organized by topic, and used by as many agents as need them. The third is the data: indexed sources copied in and prepared ahead, remote sources asked live, and clean, current documents with clear names behind all of them. The fourth is the rules: agent instructions to search, cite, and admit when it doesn't know, plus retrieval and answer instructions, and the right reasoning effort. And the fifth is testing: drills with questions you know the answers to, including ones the knowledge base can't answer, before and after every change.
The next harbour
Put together, those five lanterns turn a clever chatbot into an agent people can trust: grounded in your organization's knowledge, honest about its limits, and quick to show where each answer came from. To go further, Microsoft's documentation on Foundry IQ, on knowledge sources, and on connecting agents to knowledge bases is worth a read. Next up, you'll take an agent where people already work, by integrating it with Microsoft 365.
Facts from Microsoft's documentation: What is Foundry IQ?; What is a Knowledge Source?; Connect Agents to Foundry IQ Knowledge Bases (CC BY 4.0), rewritten in Loop's own words and examples. Narration, animation, and quiz by Loop. Microsoft, Azure, and Microsoft Foundry are trademarks of Microsoft. Loop is independent and not affiliated with or endorsed by Microsoft.