Microsoft at Build 2026: Microsoft IQ and Rayfin target agent data silos

Every new AI agent your team deploys starts from scratch. No memory of how the business works, where data lives, or what rules apply. And as agentic coding tools churn out applications faster than governance teams can keep up, those apps can become fresh silos outside your data layer. Microsoft is tackling both halves of that problem head-on at Build 2026 with Microsoft IQ and Rayfin, turning the enterprise “agent” story from a collection of isolated experiments into something more like a governed production system.

The clearest way to see the shift is what Microsoft is building for context and where it routes applications. Microsoft IQ expands Fabric IQ, its existing business data context layer, into a broader unified system called Microsoft IQ, adding three more context sources so any agent can integrate once and use a single foundation. Meanwhile, Rayfin routes agent-built applications directly to Fabric, so application data lands in Microsoft OneLake by default and feeds back into Microsoft IQ instead of accumulating somewhere else.

The timing matters. VentureBeat points to its VB Pulse Q1 2026 RAG Infrastructure Market Tracker: hybrid retrieval intent among 100-plus employee organizations tripled from 10.3% in January to 33.3% in March. In plain English, more large orgs are moving beyond basic retrieval coverage and focusing on the architecture underneath retrieval. But “shared business context” is the part retrieval alone does not solve. It is the difference between finding relevant documents and actually understanding how the organization operates: the rules, entities, workflows, and current state that shape what is “true” inside the company.

That is why Microsoft IQ is presented as an agent foundation built from four context sources that until now existed separately. Work IQ captures how the organization operates day to day, drawing on email, documents, meetings, and schedules to give agents an understanding of people, teams, and workflows. Foundry IQ manages institutional knowledge by curating and indexing knowledge bases so agents grasp what rules apply and what procedures to follow. Fabric IQ models the live operational state of the business through data, defining entities, relationships, and business rules grounded in real-time signals from Fabric Real-Time Intelligence, with ontologies expected to reach GA in the coming months. Web IQ adds real-time global context from the web, giving agents a current picture of the world outside the organization alongside internal data.

The “silos” problem does not stop at understanding. It shows up when agents start generating applications. Every new app needs a backend. Without a governed deployment path, each generated application can spawn its own isolated storage and governance pattern, drifting away from your intended data architecture. Rayfin is designed to close that loop by giving agent-built applications an enterprise-grade back end and deploying them directly to Fabric. Microsoft positions Rayfin against Supabase and Neon, Postgres-compatible backends that agentic coding tools default to, and the differentiator is governance: Rayfin routes the entire application fleet through Fabric’s unified data and compliance layer rather than letting teams proliferate disconnected silos.

Microsoft also frames the relationship between the agent and the data layer as bidirectional. Amir Netz, CTO of Microsoft Fabric, described how the platform fits using a film analogy from The Matrix: the green screen of cascading code was not just atmosphere, it was the layer that built the world. Netz told VentureBeat, “Our job in the world of data is creating reality for agents based on data.” In Microsoft’s telling, an agent building a Rayfin application draws from the organization’s ontology, and the application’s generated data then enriches that ontology for the next agent. That is an execution model shift: shared context is not only something you read, it is something the system grows over time as new applications produce new signals.

Microsoft is not the only player pushing “agent context.” VentureBeat notes that Snowflake announced context capabilities with semantic capabilities, Pinecone has its Nexus platform expanding a vector database into a knowledge engine, and Redis has developed its Iris context and memory platform. The important nuance in Microsoft’s approach is the move from RAG availability to operational trust. As Robert Kramer, managing partner at KramerERP, told VentureBeat, the enterprise AI challenge is no longer just about model availability, adding that the real question is whether Microsoft simplifies execution and strengthens trust or adds another layer to an already complex environment.

For executives and boards, the second-order implication is simple but high-stakes: the bottleneck is shifting from “Can we build an agent?” to “Can we safely scale an agent ecosystem?” If Microsoft IQ and Rayfin work as advertised, they aim to reduce the creation of new data silos by making integration and deployment follow the same shared foundation. That is exactly the kind of operational alignment enterprises typically need to move from pilots to production, especially when governance, compliance, and architecture decisions are what determine whether agent adoption becomes a strategic advantage or an expensive mess.