Ignite 2026: Microsoft Introduces IQ to Provide Shared Context for AI Agents

Microsoft IQ is designed to give AI agents access to the knowledge, data, workflows, and business context that make each organization unique. For enterprises, the platform could help reduce duplicated AI development efforts and improve consistency across AI applications, but it also requires strong governance, data management, and permissions controls to ensure agents operate on accurate and trusted information.

Cloud Computing and Security

Key Takeaways:

  • Microsoft IQ combines business data, documents, workflows, and enterprise knowledge into a shared context layer for AI agents.
  • The platform integrates Work IQ, Fabric IQ, Foundry IQ, and Web IQ to provide context across Microsoft and third-party AI ecosystems.
  • Organizations will need strong governance, permissions management, and knowledge curation to realize the benefits of enterprise-wide AI context.

As AI models become increasingly commoditized, many organizations are facing a new challenge how to make AI systems understand the unique knowledge, processes, and data that differentiate their business. Companies can access the same foundation models from providers such as OpenAI, Anthropic, and Microsoft, but competitive advantage increasingly depends on the proprietary information surrounding those models, including institutional knowledge, business processes, governance rules, and operational data.

Microsoft has introduced Microsoft IQ, which is a new enterprise context management platform designed to help organizations make their knowledge and business intelligence available to AI agents across Microsoft and third-party ecosystems. This platform brings together capabilities from Work IQ, Fabric IQ, Foundry IQ, and Web IQ, which creates a shared context layer that can be reused across AI assistants, applications, and business workflows.

Microsoft IQ creates a shared intelligence layer for AI

Microsoft IQ is an enterprise context management platform that combines data, institutional knowledge, workflows, skills, and tools into a unified intelligence layer that agents can use to understand how an organization operates.

This platform incorporates information from four major context sources. Work IQ captures documents, meetings, conversations, collaboration history, and organizational decision-making. Moreover, Fabric IQ provides governed business data, semantic models, business entities, and relationships. Foundry IQ also organizes institutional knowledge, policies, and authoritative enterprise content. Web IQ supplies external context from the web, including news, images, videos, places, and weather information.

Microsoft also says the platform supports more than 350 third-party connectors, which allows organizations to incorporate data from CRM platforms, business systems, and external services into a common intelligence framework.

Microsoft has introduced business plugins, which allow organizations to package business processes, governance policies, skills, and approved information sources into reusable components that can be shared across AI agents. These plugins can be used within Microsoft Copilot, GitHub Copilot, Microsoft Foundry, and third-party ecosystems including ChatGPT and Claude.

Microsoft expands capabilities across the four IQ platforms

Microsoft has also announced various enhancements across the platform’s four intelligence layers.

Work IQ

Microsoft is expanding Work IQ beyond information retrieval and into workflow execution. The new generally available capabilities include:

  • Business Applications in Work IQ, which adds CRM and ERP records to agent grounding.
  • Retrieve, which returns permission-aware Microsoft 365 and connected data sources through a single query.
  • Real-time meeting context for in-progress meeting intelligence.
  • Long-running tasks and agent-to-agent collaboration.
  • Granular MCP policies for access governance.
  • Support for OpenAI and Anthropic plugins.

Microsoft also announced public previews for Memory, Word, Excel, and PowerPoint creation and editing, Listen event subscriptions, and agent policy governance with human approval controls.

Fabric IQ

For Fabric IQ, Microsoft is extending governed business data into more user experiences. The new updates include:

  • Fabric data agents coming to Microsoft Copilot.
  • Fabric IQ in Excel’s Edit with Copilot experience.
  • Fabric IQ connectors for ChatGPT and Claude.
  • Fabric IQ ontologies in GitHub Copilot.
  • Fabric apps for embedding AI-powered conversational experiences into custom applications.
  • Ontology sharing capabilities through IQ Sharing.

These features are intended to make governed business context available wherever employees and AI agents work.

Foundry IQ

Microsoft is also introducing new intelligence and retrieval capabilities within Foundry IQ. The list of new capabilities includes:

  • Knowledge graphs in preview.
  • The preview release of Microsoft Retriever 1.
  • General availability of indexed SharePoint knowledge sources with permission support.
  • Foundry IQ Serverless, arriving in December 2026, with consumption-based pricing and no infrastructure management requirements.

Microsoft says these additions are designed to improve agent accuracy while reducing the amount of context and tokens required to retrieve enterprise knowledge.

Web IQ

Microsoft announced that Web IQ is now generally available to all Azure customers, with pricing starting at $1 per 1,000 requests. This service now supports unified retrieval across web content, news, images, videos, weather, and the newly added Places data source. Other capabilities include:

  • Auto and Auto Advanced retrieval modes.
  • Enterprise Scopes for domain-level inclusion and exclusion controls.
  • Bring Your Own License support for licensed publisher content.
  • Commerce data, currently in private preview, spanning more than 40 billion listings and offers.

Microsoft says Web IQ now provides access to intelligence across more than 207 million real-world entities while simplifying web grounding for AI applications.

Continuous improvement becomes a core AI strategy

Microsoft is also focusing on helping organizations continuously improve AI systems after deployment. The company introduced new optimization capabilities within Microsoft Foundry, including Agent Optimizer, which evaluates agents against business-defined success criteria and recommends improvements based on production performance.

Microsoft also announced Managed Training for automated model improvement and Programmed Training, now in public preview, which allows AI engineers to create custom training workflows such as supervised fine-tuning and reinforcement learning while Microsoft manages the underlying infrastructure.

What does this mean for enterprises?

For enterprise IT leaders, Microsoft’s announcement reflects a growing recognition that AI value depends more on the business intelligence surrounding it. Microsoft is creating a reusable context layer that can span documents, business data, enterprise knowledge, and external information sources to make agents understand how a specific organization operates.

Organizations could reduce duplicate AI development efforts, improve agent consistency, enforce governance more centrally, and accelerate deployment across departments. Support for ecosystems such as ChatGPT, Claude, GitHub Copilot, and Microsoft Copilot also suggests Microsoft is positioning IQ as a cross-platform context layer.

However, it’s important to note that implementing an enterprise context strategy will require substantial groundwork. Organizations must curate data sources, define governance policies, establish authoritative knowledge repositories, manage permissions, and determine which business processes should be exposed to agents.