Stop building RAG pipelines from scratch. Foundry IQ is the agent-native knowledge layer behind your Foundry agents. Every source in one knowledge base - Work IQ, Fabric IQ, and now Web IQ. Serverless grounding, zero friction. Available today.

@Microsoft Balanced reporting and factual analysis always stand out in today’s media landscape.

Appreciate that, Khadija. Clear, grounded reporting matters more than ever right now. #MicrosoftBuild

Thank you. In times like these, accurate, balanced, and fact-based reporting is essential. Staying grounded in the facts helps people better understand complex situations and make informed decisions.

How does Web IQ handle noisy HTML without pre chunking

Appreciate the callout. IQ now includes layout-aware ingestion, semantic ranking, and Azure Content Understanding chunking, so more of the parsing and retrieval work around noisy content is handled in-platform instead of through fully custom pipelines.

the 'stop building RAG from scratch' take is right. but for domain-specific knowledge. CRM workflows, billing logic, product docs, the answer is curated, not assembled. been exploring the hubspot navigator on for that exact slice.

Agreed. Better retrieval helps, but domain-specific knowledge still needs structure and business context to produce reliable answers.

not sure 'zero friction' holds past the setup part. when i tried grounding in my finance app the pipeline wasn't the bottleneck, retrieval quality on real data was. what does Web IQ actually do about chunking on messy web sources or is that still on you?

Fair pushback. Web IQ is meant to reduce the custom retrieval and grounding engineering teams usually build for web data, not just add another source. Retrieval quality still matters a ton though, especially with noisy finance datasets.

rip to all the custom python scripts i wrote for chunking