Microsoft introduces IQ platform to give enterprise agents shared context

Microsoft has introduced Microsoft IQ, an enterprise context management platform designed to help organizations apply their own data, processes, knowledge and tools across AI agents. The platform combines Work IQ, Fabric IQ, Foundry IQ and Web IQ, alongside more than 350 third-party connectors.
- Microsoft IQ combines Work IQ, Fabric IQ, Foundry IQ and Web IQ.
- Business plugins let organizations govern and reuse context across agents.
- The platform supports Microsoft, OpenAI and Anthropic ecosystems.
- Web IQ is available to every Azure customer with pricing starting at $1 per 1,000 calls.
- Foundry IQ Serverless is scheduled for general availability in December 2026.
A shared intelligence layer for agents
Microsoft says the advantage of an organization’s AI systems comes from the context surrounding a model. That context includes how teams work, business processes, institutional knowledge and data held in systems of record such as customer relationship management systems.
Microsoft IQ is intended to help teams curate and activate that information, as well as the skills and tools agents need. The platform brings together capabilities from Microsoft Fabric, Microsoft Foundry and GitHub to build, operate and continuously improve the intelligence used by agents.
Microsoft describes the platform as open and model-diverse for heterogeneous enterprise environments. Improvements made for one business process can become reusable assets for later updates, according to the company.
Business plugins carry governed context between agents
A central feature of Microsoft IQ is business plugins. These allow organizations to curate and govern the context, skills and authoritative sources that agents inherit. Teams can use them to capture business processes and know-how in agentic workflows, then reuse one governed plugin across agents, teams and platforms.
Microsoft gives several examples: a manufacturer could support predictive maintenance and quality root-cause analysis, an energy operator could ground control-room incident investigations, and a supply chain team could respond more quickly to supplier risk and disruption.
Users can build business plugins for agents running in Microsoft Copilot, GitHub Copilot and Microsoft Foundry through Model Context Protocol. Microsoft also lists third-party assistants, including ChatGPT and Claude marketplaces, as supported destinations.
Four IQs cover work, business data, knowledge and the web
Work IQ provides context about how people work, including documents, conversations, collaboration and past decisions. Microsoft says its Business Applications capability is now generally available, adding customer relationship management and enterprise resource planning records to the grounding set. Retrieve is also generally available, returning permission-trimmed Microsoft 365 content and federated connector data in one call.
Fabric IQ provides governed business data, entities, relationships and rules. Power BI semantic models in Microsoft Copilot are generally available, while Fabric data agents are coming to Microsoft Copilot, with general availability coming soon. Fabric IQ connectors for ChatGPT and Claude are in preview.
Foundry IQ focuses on institutional knowledge. Knowledge graphs and Microsoft Retriever 1 are in preview, while the SharePoint indexed knowledge source is generally available with permissions support. Foundry IQ Serverless is scheduled for general availability in December 2026, offering usage-based pricing without capacity to provision or manage.
Web IQ is available to every Azure customer with no waitlist. Microsoft says pricing starts at $1 per 1,000 calls. The service adds Auto and Auto Advanced modes and brings together web, images, videos, news, places and weather. It also includes controls for specific sites and domains, while Commerce is in private preview.
For users and businesses, Microsoft IQ is intended to reduce the need to repeatedly explain organizational rules and processes when creating agents. Microsoft points to AEMO’s NEM-GEM, an AI-powered operational intelligence capability co-engineered with Microsoft in less than 12 weeks, as an example of the platform being used to help grid operators focus on important signals while humans remain in control.

