If you have spent any time reading about Microsoft's agentic AI announcements, you will have run into four similar-sounding names - Work IQ, Fabric IQ, Foundry IQ and Web IQ - and quite possibly come away unsure which is which. They are easy to confuse, and most explanations bury the distinction under jargon.

This guide fixes that. The simplest way to keep the four straight is to remember that each one answers a different question an AI agent needs answered before it can act reliably. Get the questions clear and the rest falls into place.

First, the umbrella: Microsoft IQ

All four sit under one idea: Microsoft IQ, the "context layer" that grounds AI agents in your organisation's real knowledge rather than leaving them to guess. A capable model knows language; it does not know your business. Microsoft IQ is what supplies that missing context, and it works across Microsoft's agent surfaces including GitHub Copilot, Microsoft Foundry and Copilot Studio.

Think of the model as the brain and Microsoft IQ as the memory and context that brain needs to be useful inside your organisation. The four "IQs" are simply four different kinds of context - each drawing on a different source and answering a different question.

The four layers at a glance

The sections below take each one a little deeper.

Work IQ - "how work happens here"

Work IQ draws on the signals inside Microsoft 365: the people, documents, emails, meetings and chats across your organisation, and - crucially - the relationships between them. It is what lets an agent understand that a particular person sits on a particular team, that the relevant document lives in a particular place, and who tends to work on what.

Because it operates inside the Microsoft 365 trust boundary, Work IQ respects existing permissions: an agent grounded in Work IQ should only see what the user it is acting for is already allowed to see. Reach for it when your agent's usefulness depends on understanding the human and content fabric of how your organisation actually works.

Fabric IQ - "what the business means"

Fabric IQ is the shared layer of business meaning over your structured data. It is where the semantics live - the agreed definitions of your core business concepts (what counts as an "active customer", when an order is "complete") and how those concepts relate to one another, increasingly expressed as ontologies. It builds on your data foundation in OneLake and on the semantic models your analytics teams already maintain.

This is the layer that stops two agents giving two different answers to the same business question. But it comes with a dependency worth stating plainly: a semantic layer is only as trustworthy as the data and definitions beneath it. If your underlying data is duplicated or your business terms are not agreed, Fabric IQ will faithfully serve those inconsistencies to every agent. Getting that foundation right - quality, matching and governance, the work our MatchX platform is built for - is what makes Fabric IQ deliver. Reach for it when your agents need a consistent, governed understanding of your business's structured data.

Foundry IQ - "what we already know"

Foundry IQ is about knowledge and retrieval. It unifies different knowledge sources behind a single retrieval endpoint so that an agent can discover and reuse knowledge rather than starting from scratch each time, and so that the same knowledge can serve many agents across many deployments.

For organisations sitting on large volumes of documents, policies and unstructured content, this is the layer that turns that material into something agents can actually use - which is precisely the problem our PromptX platform addresses, turning documents into governed, retrievable knowledge. Reach for it when your agents need to find and reuse organisational knowledge reliably and consistently.

Web IQ - "what's true in the world right now"

The newest of the four, Web IQ adds live, real-time context from the open web. It exists because an agent grounded only in internal data and a fixed training cut-off has a blind spot for anything current - market movements, published guidance, external events. Web IQ is the grounding system that fills that gap with up-to-date, web-scale information. Reach for it when your agent needs to reason about the present-day world, not just your internal estate.

How they work together?

In practice these layers combine. Imagine an agent helping a relationship manager prepare for a client meeting. Work IQ tells it who the client contacts are and surfaces the relevant internal threads and documents. Fabric IQ gives it the agreed definition of that client's account status and the right figures behind it. Foundry IQ retrieves the relevant policy and product knowledge. Web IQ adds anything material that has happened in the client's market this week. No single layer would produce a good briefing; together they produce a grounded one.

That is the real point of Microsoft IQ: context is becoming the differentiator in enterprise AI. Two organisations using the same model get very different results depending on how well their agents are grounded - and these four layers are how that grounding is delivered.

What this means for you?

The four IQs are a powerful way to deliver context to agents. They do not create good context out of poor inputs. Work IQ inherits your M365 permissions; Fabric IQ inherits your data quality and definitions; Foundry IQ inherits the state of your knowledge; even Web IQ needs your agents configured to use it sensibly. The organisations that get the most from the context layer are the ones whose foundations are ready - data unified, matched and governed, knowledge organised, and definitions agreed.

That readiness is worth checking before you build. Our Agent-Ready Data checklist is a quick way to see whether your foundation can support the context these layers are designed to deliver, and our guide to what Microsoft IQ is gives the bigger picture if you want to start there.

VE3 is a Microsoft solution partner helping regulated organisations adopt the context layer on solid data and knowledge foundations. Start a readiness conversation.