Microsoft has announced Microsoft IQ, a platform intended to give AI agents reusable context about an organisation’s people, data, processes and knowledge. The practical pitch is that this context can travel between agents and services, instead of being rebuilt for every tool.
Microsoft AI Watch analysis
What happened
Microsoft IQ combines four parts: Work IQ for how people work, Fabric IQ for governed business data, Foundry IQ for institutional knowledge, and Web IQ for current information from the web. Microsoft says the platform includes more than 350 third-party connectors and is designed to work across different models and services.
A central feature is the new business plugins, which let teams package and govern context, skills and authoritative sources for reuse across agents and platforms, including Copilot, GitHub Copilot, ChatGPT and Claude. Microsoft also says several Work IQ capabilities are generally available, including CRM and ERP records as grounding sources, permission-trimmed retrieval and real-time meeting context. Other features, such as agent policy governance with human approval and Office document creation and editing, are in public preview.
Why it matters
The hard part of making an AI agent useful at work is often not finding a model. It is giving the model the right company information, while respecting who is allowed to see or do what. Microsoft IQ’s bet is that organisations can curate that context once and reuse it across agents, rather than building isolated integrations for each one.
That could make agents more useful across real workflows, from searching company knowledge to acting on business data. It also puts governance in the product pitch: Microsoft says granular MCP policies can scope access by path, audience and trusted app. A platform announcement is not proof that these controls work well in every deployment, but it does give buyers specific capabilities and availability claims to check.
Our read
This is a substantial attempt to make enterprise context a reusable layer for AI, not just another model announcement with a new badge. The most meaningful test will be whether organisations can carry useful context between tools without making permissions harder to understand. For teams evaluating it, the distinction between generally available features and public previews matters more than the IQ naming scheme, which is already doing enough heavy lifting.
What to watch
- Which Microsoft IQ capabilities move from preview to general availability, and when.
- Whether business plugins work across the listed Microsoft and third-party agent platforms.
- How administrators can inspect and manage permissions as context is reused across agents.
Discussion spark: Would you trust a shared, reusable company context layer to make AI agents more useful, or does centralising that context create a governance problem of its own?
Sources and evidence
- Microsoft's AI pitch at Ignite: Cheaper models as business 'token spend is going through the roof' – AOL.co.uk (17 November 2026, 20:34 UTC)
- Introducing the new Microsoft IQ: Putting your organization's intelligence to work in every agent – Microsoft Azure (17 November 2026, 20:34 UTC)
- Microsoft and Oracle turn data into insights and actions – Microsoft Azure (17 November 2026, 20:34 UTC)
not affiliated with or endorsed by Microsoft