Foundry IQ for Multi-Source AI Knowledge Bases

8 snips

Dec 3, 2025

In this engaging discussion, Pablo Castro, a Distinguished Engineer and CVP for Azure AI Search, delves into the intricacies of creating powerful AI agents that pull knowledge from various sources. He explains how Foundry IQ automates connections across complex data landscapes, ensuring accurate and context-rich answers. Listeners learn about reusing existing Azure AI Search assets, optimizing retrieval efforts, and the innovative architecture behind multi-source knowledge bases. Pablo’s insights promise smarter AI solutions with less engineering hassle!

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Multi-Source Grounding Works Seamlessly

Foundry IQ lets agents pull data from many scattered enterprise sources and connect the dots automatically.
This removes manual orchestration and enables agents to ground answers across lakes, SharePoint, and the web.

ANECDOTE

Technician Agent Demo Uses Three Sources

Pablo demos a technician support agent that consulted three sources to answer a messy on-site question.
The agent combined machine manuals, SharePoint policy, and public web standards to produce a grounded reply.

INSIGHT

Knowledge Bases Encapsulate Planning And Sources

Knowledge bases in Foundry IQ encapsulate an information domain and its sources plus retrieval models.
The system plans query decomposition, runs searches, iterates if needed, and returns synthesized answers with citations.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Ground your AI agents on your organization's knowledge, pulling from multiple sources at once, connecting the dots automatically, and getting more accurate, context-rich answers without doing manual orchestration with Foundry IQ in Microsoft Foundry. Navigate complex, distributed data across Azure stores, SharePoint, OneLake, MCP servers, and even the web, all through a single knowledge base that handles query planning and iteration for you.

Reuse the Azure AI Search assets you already have, build new knowledge bases with minimal setup, and control how much reasoning effort your agents apply. As you develop, you can rely on iterative retrieval only when it improves results, saving time, tokens, and development complexity.

Pablo Castro, Azure AI Search CVP and Distinguished Engineer, joins Jeremy Chapman to share how to build smarter, more capable AI agents, with higher-quality grounded answers and less engineering overhead.

► Link References

To learn more check out https://aka.ms/FoundryIQ

For more details on the evaluation metric discussed on this show, read our blog at https://aka.ms/kb-evals

For more on Microsoft Foundry go to https://ai.azure.com/nextgen

► Unfamiliar with Microsoft Mechanics?

As Microsoft's official video series for IT, you can watch and share valuable content and demos of current and upcoming tech from the people who build it at Microsoft.

• Subscribe to our YouTube: https://www.youtube.com/c/MicrosoftMechanicsSeries

• Talk with other IT Pros, join us on the Microsoft Tech Community: https://techcommunity.microsoft.com/t5/microsoft-mechanics-blog/bg-p/MicrosoftMechanicsBlog

• Watch or listen from anywhere, subscribe to our podcast: https://microsoftmechanics.libsyn.com/podcast

► Keep getting this insider knowledge, join us on social:

• Follow us on Twitter: https://twitter.com/MSFTMechanics

• Share knowledge on LinkedIn: https://www.linkedin.com/company/microsoft-mechanics/ • Enjoy us on Instagram: https://www.instagram.com/msftmechanics/

• Loosen up with us on TikTok: https://www.tiktok.com/@msftmechanics

► QUICK LINKS:

00:00 - Foundry IQ in Microsoft Foundry

01:02 - How it's evolved

03:02 - Knowledge bases in Foundry IQ

04:37 - Azure AI Search and retrieval stack

05:51 - How it works

06:52 - Visualization tool demo

08:07 - Build a knowledge base

10:10 - Evaluating results

13:11 - Wrap up