Back home

Microsoft IQ

x+web 0.95
Visit source

Intro

Microsoft IQ is the enterprise context management platform Microsoft announced at Ignite on 17 November 2026. The Azure Blog post Introducing the new Microsoft IQ starts from a simple claim: everyone can use the same frontier models, and those models improve every few weeks. What sets an organization apart is everything around the model—how its people work, its processes, its institutional knowledge, and data in systems of record such as CRM. Microsoft IQ is meant to help an organization curate that intelligence and put it to work in its agents.

Officially it unifies four complementary pieces: Work IQ for how people work (documents, conversations, collaboration, and past decisions), Fabric IQ for governed business data, entities, relationships, and rules, Foundry IQ for curated institutional knowledge such as policies and authoritative documents, and Web IQ for timely context from the web, plus more than 350 third-party connectors. New business plugins let teams capture context, skills, and authoritative sources so agents can inherit them in Microsoft Copilot, GitHub Copilot, Microsoft Foundry via MCP, and the ChatGPT and Claude marketplaces.

This entry records the umbrella product name Microsoft IQ as announced and made generally available. It is not the same name as any one pre-existing Work IQ or Fabric IQ feature, and it should not be confused with unrelated model names such as IQuest-Q1.

Advertisement

Explain

Microsoft’s argument is that models themselves are commoditizing quickly, so an organization’s advantage depends on whether its own context can be used by agents in a stable, governed way. Microsoft IQ gathers context that previously lived across Microsoft 365, Fabric, Foundry, and the web, and uses business plugins to freeze “which knowledge, skills, tools, and sources an agent should inherit,” so the next task does not start from a blank setup.

The same post also lists component-level updates: in Work IQ, Business Applications for CRM/ERP, Retrieve, real-time meeting context, long-running tasks, and agent-to-agent support are called generally available; Foundry IQ adds knowledge graphs in preview and a retrieval model in preview named Microsoft Retriever 1; Web IQ is available to every Azure customer, with usage-based pricing starting at a published per-call rate. Those are capabilities under the same platform; this entry does not split them into separate terms unless they later receive a standalone product launch.

Microsoft IQ is not a new general-purpose foundation model. It is a platform layer that feeds organizational context to models and agents. It should also not be merged with other Ignite names such as Microsoft Perception, Copilot Home, or the MAI model family.

Usage

Start with the Azure Blog announcement, then use Microsoft’s documentation to enable context, connectors, and business plugins in Copilot, GitHub Copilot, Foundry, or third-party assistants over MCP. The intended use is to curate the knowledge, processes, and tools an organization wants agents to inherit, package them as governed plugins, and reuse that same context across teams instead of wiring each app to its data sources from scratch.

Evaluate which systems it can actually reach, whether permissions are trimmed per person, whether plugins survive a model swap, and how metered pieces such as Web IQ or Foundry IQ Serverless add up. Licensing, regions, and anything still in preview follow the admin experience and official docs at the time you turn it on.

If you only ask Copilot a question about a company document, or only use a Fabric semantic model, that is not this new umbrella product name itself.