Work IQ Explained: Giving AI Agents Context for How Work Gets Done

An AI agent can search an inbox, open a project document, read a Teams conversation, and review a calendar.

But access to workplace data does not automatically mean the agent understands the work.

To prepare an accurate project update, the agent may need to know which document is current, what decisions were made in the last meeting, who owns each action item, which deadline matters, and whether a recent email changed the plan. It may also need to understand how the people, conversations, files, meetings, and tasks relate to one another.

That is where Microsoft Work IQ becomes important.

Work IQ gives Microsoft 365 Copilot and other AI agents a continuously updated understanding of how work happens across an organization. It connects workplace data with relationships, preferences, work patterns, tools, and permissions so agents can provide answers and take actions that reflect the user’s actual working context.

Microsoft describes Work IQ as a workplace intelligence layer that allows agents to access and reason over organizational data, context, and tools. Instead of requiring developers to build separate retrieval systems for email, meetings, documents, people, and collaboration data, Work IQ assembles that context into an agent-ready intelligence layer.

Quick Answer

Work IQ is the intelligence layer behind Microsoft 365 Copilot and Microsoft’s broader agent ecosystem.

It combines signals from emails, meetings, documents, chats, people, organizational relationships, work patterns, and connected business systems. Work IQ then uses that context to help agents understand what a user is working on, find relevant information, select appropriate tools, maintain continuity across tasks, and take governed actions.

Microsoft Graph provides access to workplace data. Work IQ builds on that foundation by interpreting the relationships and context surrounding the data so agents can reason about how work gets done.

In This Blog

• What Is Microsoft Work IQ?

• Why AI Agents Need More Than Workplace Data

• What Does “How Work Gets Done” Actually Mean?

• The Three Intelligence Layers of Work IQ

• The Four Building Blocks of the Work IQ API

• How Microsoft Graph Supports Work IQ

• How Semantic Indexing Improves Agent Context

• Extending Work IQ Beyond Microsoft 365

• How AI Agents Use Work IQ

• Work IQ Versus Connectors and Traditional RAG

• How Work IQ Fits into Microsoft IQ

• Security, Permissions, and Governance

• A Practical Work IQ Example

• How Organizations Can Prepare for Work IQ

• What Work IQ Means for the Enterprise

• Frequently Asked Questions

What Is Microsoft Work IQ?

Work IQ is Microsoft’s workplace intelligence layer for Copilot, agents, applications, and agentic workflows.

It brings together Microsoft 365 work data, including emails, meetings, documents, chats, calendars, people, and organizational information. It then combines those signals with contextual information such as relationships, preferences, priorities, and work patterns.

According to Microsoft’s Microsoft 365 Copilot extensibility overview, Work IQ uses this intelligence to:

• Assemble relevant context

• Ground agent responses

• Select appropriate skills

• Invoke tools

• Maintain continuity across interactions

• Support actions within Microsoft 365

This makes Work IQ more than a data connector.

A connector can give an agent access to an email or document. Work IQ helps the agent determine which email or document is relevant to the current task, how it relates to other workplace activity, and what the user may need to do next.

Microsoft’s broader vision is to make this context reusable across Microsoft 365 Copilot, Copilot Studio, Microsoft Foundry, Agent 365, custom applications, and third-party agents. The same intelligence layer that helps Copilot understand a user’s work can therefore support agents built for specific roles, processes, or industries.

Microsoft made the Work IQ API generally available on June 16, 2026. However, availability, licensing, billing, and feature status can vary across Microsoft products and integration surfaces. For example, Microsoft’s current Copilot Studio documentation may still identify some Work IQ MCP capabilities as preview features. Organizations should confirm the latest Microsoft Work IQ release information and product documentation before planning a production deployment.

Why AI Agents Need More Than Workplace Data

Most knowledge work is distributed across applications.

A project may involve:

• A statement of work in SharePoint

• Requirements discussed in Teams

• Decisions recorded in meeting transcripts

• Deadlines on several calendars

• Action items in an email thread

• Ownership information in Microsoft Entra ID

• Customer records in a CRM system

• Supporting data in an external application

An agent could technically connect to all of those sources and still produce an incomplete or misleading response.

For example, suppose a user asks:

What do I need to know before tomorrow’s customer meeting?

A basic search system may find documents and messages containing the customer’s name. However, a useful agent needs to go further.

It should identify:

• The upcoming meeting on the user’s calendar

• The attendees and their roles

• The latest customer communications

• The current version of the project plan

• Unresolved commitments from previous meetings

• Recent changes that may affect the discussion

• Files the user is authorized to access

• Potential next steps the user can take

That requires context, not simply keyword matching.

Microsoft explains in its Work IQ documentation that Work IQ continuously builds a semantic understanding across Microsoft 365 and connected external systems. This allows agents to reason over organizational information without requiring developers to manually combine raw signals or create a separate retrieval pipeline for every scenario.

What Does “How Work Gets Done” Actually Mean?

The phrase “how work gets done” includes more than the content stored in Microsoft 365.

It also includes the relationships and activity surrounding that content.

People

Who created the document? Who owns the project? Who reports to whom? Who has expertise in a specific subject? Who regularly participates in decisions about this account?

Content

Which file is the approved version? Which documents are commonly referenced together? Which presentation was used in the most recent executive review?

Communication

What was discussed in email, Teams, and meetings? Did the customer approve the proposal? Was a decision changed later? Are there unresolved questions?

Time

Which meeting is upcoming? Which deadline is approaching? When was the information last updated? Is an earlier decision still current?

Relationships

How does a person relate to a project, meeting, customer, team, or document? Which conversations and files belong to the same body of work?

Work Patterns and Preferences

Which sources does the user typically rely on? Which teammates are normally involved? What level of detail does the user prefer? Which actions usually follow a particular event?

These signals help an agent interpret workplace information as part of an active process rather than as isolated files and messages.

Microsoft says Work IQ combines Microsoft 365 work data with relationships, preferences, and work patterns. That context helps agents ground responses, choose skills, and invoke tools based on the current user and task. Learn more in Microsoft’s Copilot extensibility documentation.

The Three Intelligence Layers of Work IQ

Microsoft describes Work IQ through three tightly connected intelligence layers: Data, Memory, and Inference.

These layers explain how workplace signals become useful agent context.

1. Data

The data layer brings together signals from:

• Files

• Email

• Meetings

• Calendars

• Microsoft Teams

• People and organizational profiles

• SharePoint Lists

• Microsoft 365 Search

• Connected line-of-business systems

The purpose is not simply to collect more content. It is to create a current view of the information and activity surrounding the user’s work.

Microsoft’s current Work IQ sources documentation identifies mail, calendars, Teams, OneDrive, SharePoint, user profiles, people information, Microsoft 365 Copilot Search, and SharePoint Lists as available Work IQ source categories.

2. Memory

Memory helps agents retain useful context across interactions, applications, and longer-running tasks.

For example, an agent may need to remember:

• The objective of an ongoing assignment

• The files it has already reviewed

• Intermediate conclusions

• Decisions made during previous steps

• The user’s priorities

• Which tasks remain incomplete

This does not mean an agent should indiscriminately retain everything. Memory must remain subject to organizational controls, task design, permissions, retention requirements, and the capabilities of the specific agent platform.

Microsoft explains that Work IQ memory helps Agent 365-managed agents remain aligned with priorities and stay consistent across tasks, applications, and sessions. See Microsoft’s Work IQ guidance for Copilot Studio.

3. Inference

Inference brings together models, skills, context, and tools so an agent can reason about what it has found and determine what to do next.

The agent may need to:

• Identify relevant information

• Compare different sources

• Resolve chronology

• Recognize a missing decision

• Determine which tool to use

• Draft a response

• Update a resource

• Ask a user for approval

Microsoft positions Agent 365 as the control plane surrounding these agent actions, helping organizations maintain visibility, policy enforcement, and compliance as agents reason and act.

The Four Building Blocks of the Work IQ API

The Data, Memory, and Inference layers describe the intelligence Work IQ provides.

For developers, Microsoft also organizes the Work IQ API around four functional building blocks: Chat, Context, Tools, and Workspaces.

Chat

Chat supports conversational interaction through both Agent-to-Agent protocols and REST-based application experiences.

This allows agents to exchange work, maintain task continuity, and return processed responses rather than forcing the calling application to orchestrate every intermediate step.

Context

Context is one of Work IQ’s most important capabilities.

Instead of asking developers to manually retrieve email, files, meetings, and people information, Work IQ assembles relevant organizational context internally. It then provides agent-ready inputs based on the user, request, and available information.

Tools

Tools allow agents to retrieve data and perform actions across Microsoft 365.

Microsoft says the Work IQ MCP surface consolidates hundreds of Microsoft 365 operations into 10 generic tools. These tools use simple verbs such as fetch, create, and update, while resource paths identify the Microsoft 365 resource involved.

Work IQ can also allow an agent to discover how a data source is structured at runtime. This reduces the need to hard-code a separate integration model for every resource. Learn more in Microsoft’s Work IQ developer documentation.

Workspaces

Workspaces support longer, multistep agent tasks.

Work IQ uses SharePoint Embedded working storage to give agents a persistent area inside the Microsoft 365 tenant boundary. Agents can use this space to store intermediate files, preserve task progress, reuse previous results, and hand work from one agent or experience to another.

How Microsoft Graph Supports Work IQ

Microsoft Graph is a foundational part of Work IQ, but the two are not the same.

Microsoft Graph provides APIs for accessing data and services across Microsoft 365. Depending on permissions, developers can use Graph to work with resources such as:

• Users

• Groups

• Email

• Calendars

• Files

• SharePoint sites

• Teams

• Tasks

• Devices

• Organizational information

Microsoft 365 Copilot already uses Graph to bring personalized workplace context into prompts, including information from a user’s emails, chats, documents, and meetings. Microsoft explains this process in its Microsoft 365 Copilot overview.

However, Graph primarily provides access to resources and their underlying signals.

Work IQ adds an intelligence layer over that foundation.

A simple way to understand the distinction is:

Microsoft summarizes this distinction by explaining that Graph exposes workplace data, while Microsoft IQ turns data and other signals into reusable enterprise intelligence that agents can reason over. See Microsoft’s Microsoft IQ overview.

How Semantic Indexing Improves Agent Context

Traditional enterprise search depends heavily on filenames, keywords, metadata, and exact wording.

People rarely remember information that way.

A user may ask for “the proposal we discussed before the Chicago meeting,” even though none of those exact words appear in the filename.

Microsoft’s semantic index for Copilot helps address this problem by mapping organizational data into lexical and semantic indexes. This allows Microsoft 365 experiences to retrieve information based on meaning, context, and relationships rather than relying only on direct keyword matches.

For Work IQ, semantic indexing can help an agent connect:

• A meeting to the documents discussed during it

• A user to the people they regularly collaborate with

• A project name to related shorthand or terminology

• A question to content with similar meaning

• A current task to earlier discussions

• A file to the people and activities surrounding it

The result is more relevant retrieval, but retrieval remains only one part of the process.

Work IQ must still evaluate context, permissions, chronology, relationships, available tools, and the user’s actual request before the agent can produce a useful response.

Extending Work IQ Beyond Microsoft 365

Not all workplace knowledge lives in Microsoft 365.

Organizations may also rely on:

• CRM platforms

• Human resources systems

• Customer support applications

• Project management tools

• Learning management systems

• Engineering repositories

• Document management platforms

• Industry-specific applications

Microsoft 365 Copilot connectors allow organizations to bring external information into Microsoft 365 intelligence experiences.

Microsoft currently supports two primary connector approaches.

Synced Connectors

Synced connectors ingest and index external content in Microsoft Graph.

This approach is useful for document repositories, knowledge systems, and other sources where semantically indexed search can improve discovery and synthesis.

Federated Connectors

Federated connectors retrieve information from the source at query time through Model Context Protocol.

The data is not indexed into Microsoft Graph. Instead, the agent calls the external system when current information is needed.

This approach may be appropriate for:

• Rapidly changing information

• Regulated content that should remain in its source system

• Dynamic line-of-business data

• Systems where real-time retrieval is important

Microsoft’s Copilot connectors documentation explains the differences in data movement, indexing, authentication, and retrieval between synced and federated connectors.

Microsoft also provides specialized people data connectors that can enrich Microsoft 365 with authoritative information from HR, recruiting, talent, or other people systems. This can improve an agent’s understanding of roles, skills, reporting structures, and organizational context while allowing the external system to remain the authoritative source.

How AI Agents Use Work IQ

Work IQ can support both knowledge retrieval and action-oriented agent experiences.

Finding Relevant Information

An agent could locate the latest project documents, summarize recent discussions, identify open questions, and highlight changes since the last meeting.

Preparing for Meetings

An agent could review the calendar event, attendee profiles, prior meeting notes, relevant emails, and active documents before preparing a briefing.

Drafting Communications

An agent could examine the history of a request, understand the people involved, and draft a response that reflects recent decisions and current priorities.

Supporting Project Handoffs

An agent could assemble files, decisions, action items, deadlines, owners, and unresolved risks for a new project lead.

Coordinating Work

An agent could identify available attendees, suggest meeting times, retrieve supporting materials, and prepare an agenda.

Running Longer Workflows

An agent could perform research across multiple sources, preserve interim findings in a workspace, delegate a task to another agent, and return a completed deliverable.

Acting Across Microsoft 365

Depending on the tools, permissions, and product surface, an agent may be able to create, update, or otherwise interact with Microsoft 365 resources.

Microsoft currently exposes Work IQ through several integration formats:

• Agent-to-Agent, or A2A: For agent collaboration

• Model Context Protocol, or MCP: For agent tools and runtimes

• REST APIs: For custom web and application experiences

This gives organizations options for building Work IQ into low-code agents, pro-code applications, third-party products, and multi-agent systems. See Microsoft’s Work IQ API documentation for the current integration options.

Work IQ Versus Connectors and Traditional RAG

Work IQ does not eliminate the need for connectors, search, or retrieval-augmented generation. Instead, it addresses additional layers of the problem.

A custom RAG system may be appropriate when an organization needs complete control over retrieval, embeddings, ranking, infrastructure, or a specialized data environment.

However, building that system can require:

• Custom indexes

• Vector storage

• Permissions synchronization

• Retrieval pipelines

• Data refresh processes

• Security enforcement

• Compliance controls

• Observability

• Tool orchestration

Microsoft says Work IQ allows applications to use the intelligence behind Microsoft 365 Copilot without requiring teams to build custom indexing, vector stores, or compliance pipelines for Microsoft 365 work data. Learn more in Microsoft’s Copilot extensibility overview.

The right architecture may combine these approaches. For example, an agent could use Work IQ for Microsoft 365 context, Fabric IQ for governed business data, and Foundry IQ for authoritative enterprise knowledge.

How Work IQ Fits into Microsoft IQ

Work IQ is one component of Microsoft’s broader enterprise intelligence strategy.

Microsoft IQ is designed to provide Copilot, agents, and applications with a shared understanding of the enterprise.

Its four primary components address different types of context.

Consider a customer renewal agent.

Work IQ could help it understand:

• Recent customer emails

• Upcoming account meetings

• Internal Teams discussions

• The account team

• Relevant documents

• Open commitments

Fabric IQ could provide:

• Contract value

• Product usage

• Support metrics

• Renewal probability

• Account health definitions

Foundry IQ could provide:

• Approved pricing policies

• Renewal playbooks

• Legal guidance

• Product documentation

Web IQ could provide:

• Current news about the customer

• Market developments

• Public leadership changes

• Recent industry events

Together, these context layers allow the agent to understand the work, the business, the organization’s knowledge, and the external environment.

Security, Permissions, and Governance

Work IQ can make workplace information easier for agents to find and use. Therefore, the quality of an organization’s existing Microsoft 365 governance becomes even more important.

User-Scoped Access

Work IQ requests operate in the context of a signed-in user.

Microsoft states that Work IQ respects Microsoft 365 permissions and sensitivity labels. An agent should only access information the current user is authorized to access. See Microsoft’s Microsoft 365 Copilot extensibility documentation.

Similarly, Microsoft 365 Copilot uses identity-based access controls. It does not provide a user with access to a SharePoint site, Teams channel, email, or document that the user could not otherwise access. Microsoft explains these controls in its Microsoft 365 Copilot architecture documentation.

However, this creates an important readiness issue: Copilot and agents can surface information that users technically have permission to access, even when that access is broader than the organization intended.

Organizations should review:

• Overshared SharePoint sites

• Public or broadly accessible Teams channels

• Outdated group membership

• External sharing

• Anonymous links

• Inherited permissions

• Sensitive information without labels

• Stale documents

• Content retention

• Guest access

Policy-Based Tool Controls

Work IQ also introduces centralized controls over agent tools and actions.

Microsoft says the Work IQ MCP uses broad access boundaries combined with a context-aware policy engine. Policies can evaluate factors such as:

• User identity

• Resource path

• Requested method

• Data content

• Intended action

This allows administrators to control not only what an agent can retrieve, but also what it can do with the information.

Agent 365 Governance

Agent 365 provides a control plane for agent identity, approved tools, security, and observability.

Depending on regional and product availability, administrators can use Microsoft 365 management experiences to allow or block tooling servers, scope permissions, and manage agent access.

Microsoft also describes integration with Microsoft Defender for reviewing agent tool calls, execution details, parameters, results, and potential anomalies. See Microsoft’s Work IQ and Agent 365 documentation.

Data Protection

Microsoft states that prompts, responses, and organizational data accessed through Microsoft Graph are not used to train the foundation models used by Microsoft 365 Copilot. Agents also operate within existing Microsoft 365 data access boundaries rather than introducing new user privileges. Microsoft provides additional detail in its Copilot data privacy and security documentation.

Work IQ provides the technical mechanisms for permission-aware context. Organizations still need to classify information, correct oversharing, define agent policies, monitor activity, and determine which actions require approval.

A Practical Work IQ Example

Consider a professional services firm preparing for a quarterly client review.

Information about the engagement is spread across:

• Outlook email threads

• Teams conversations

• Meeting transcripts

• SharePoint documents

• Employee calendars

• Project plans

• CRM records

• Financial reporting systems

The account lead asks an agent:

Prepare me for tomorrow’s client review. Summarize progress, decisions, open risks, outstanding commitments, and the people who need to be involved.

Without Work IQ

A conventional agent may search for the customer name and retrieve a long list of documents and messages.

However, it may struggle to determine:

• Which project plan is current

• Which decision superseded an earlier decision

• Whether an email commitment remains open

• Who owns a specific risk

• Which meeting is relevant

• Which participants have decision-making authority

• Which documents the user should review first

With Work IQ

Work IQ can help assemble the relevant workplace context:

1. The calendar identifies the upcoming client review.

2. People and organizational data identifies the attendees and their roles.

3. Outlook provides recent client communications.

4. Teams provides internal project discussions.

5. SharePoint and OneDrive provide supporting documents.

6. Meeting data provides prior decisions and action items.

7. Semantic search connects related content even when naming conventions differ.

8. Work patterns and relationships help prioritize the most relevant information.

The agent could then prepare a structured briefing that includes:

• Current project status

• Decisions since the last review

• Outstanding customer requests

• Internal risks

• Upcoming milestones

• Assigned owners

• Recommended discussion topics

• Supporting source links

With appropriate tools and permissions, the agent could also draft the meeting agenda, prepare a follow-up email, or create a workspace for the review process.

If Fabric IQ is added, the same briefing could incorporate governed project financials, utilization, margins, delivery metrics, and account health scores.

The value comes from connecting workplace context with business context rather than treating each source as an independent search result.

How Organizations Can Prepare for Work IQ

Work IQ should not be treated as a feature that IT simply turns on.

Its usefulness depends on the quality of the organization’s Microsoft 365 environment, permissions, content, people data, agent design, and governance.

1. Choose a Specific Work Scenario

Start with a task that requires context across multiple workplace sources.

Examples include:

• Meeting preparation

• Customer account handoffs

• Executive briefings

• Project status updates

• Employee onboarding

• Incident coordination

• Proposal development

• Internal research

The use case should have a clear user, outcome, and measure of success.

2. Map the Context the Agent Needs

Identify the sources, relationships, and activities required for the task.

Ask:

• Which emails matter?

• Which meetings provide decisions?

• Which SharePoint sites contain authoritative documents?

• Which people or roles are involved?

• Which business systems provide supporting data?

• What should the agent be allowed to create or update?

This prevents the organization from connecting broad sources without understanding how they contribute to the actual task.

3. Review Microsoft 365 Permissions

Before expanding agent access, evaluate SharePoint, Teams, OneDrive, group membership, guest access, and external sharing.

Permission-based grounding is only as precise as the permissions already in place.

4. Improve Content Quality

Agents will have difficulty identifying authoritative information when an environment contains:

• Duplicate files

• Poor naming conventions

• Obsolete policies

• Missing ownership

• Inconsistent metadata

• Multiple versions labeled “final”

• Abandoned SharePoint sites

Content lifecycle management becomes part of AI readiness.

5. Strengthen People and Organizational Data

Agents need accurate information about roles, departments, reporting structures, skills, and organizational relationships.

Review Microsoft Entra ID, Microsoft 365 profiles, and connected HR sources. Where necessary, use people data connectors to enrich Microsoft 365 with authoritative profile information.

6. Decide How External Systems Should Connect

Use synced connectors for content that benefits from semantic indexing and broad discovery.

Use federated MCP-based connections when the information must remain in the source system or needs to be retrieved in real time.

7. Define Agent Permissions and Actions

Separate knowledge access from action permissions.

An agent that can read calendar availability presents a different level of risk than an agent that can create meetings, send email, or update business records.

Start with the minimum tools needed for the scenario.

8. Establish Human Approval Points

High-impact actions should not automatically occur simply because an agent can perform them.

Consider requiring approval before:

• Sending external communications

• Changing deadlines

• Updating customer records

• Sharing documents

• Creating financial commitments

• Communicating sensitive information

• Triggering operational workflows

9. Test Realistic Questions

Test with the language employees actually use, including incomplete or ambiguous requests.

Evaluate:

• Relevance

• Source accuracy

• Permission enforcement

• Chronological reasoning

• Recognition of authoritative content

• Understanding of people and roles

• Appropriate tool selection

• Handling of conflicting information

• Citation quality

• Escalation behavior

10. Measure Business Results

Useful measures may include:

• Time saved preparing for meetings

• Reduction in manual searching

• Faster project handoffs

• Fewer missed commitments

• Better response consistency

• Improved employee onboarding

• Shorter decision cycles

• Reduced administrative work

The goal is not to prove that an agent can retrieve information. It is to show that the agent can improve how real work gets done.

What Work IQ Means for the Enterprise

Enterprise AI often begins with a model, a prompt, and access to a knowledge source.

That may be enough for a narrow question-answering tool. It is not enough for agents expected to participate in real work.

An effective workplace agent must understand:

• Who the user is

• What the user is working on

• Which information is current

• How people and content relate

• Which decisions have been made

• What remains incomplete

• Which tools are available

• What the user is permitted to do

• Which actions require oversight

Work IQ provides Microsoft’s intelligence layer for that workplace context.

Microsoft Graph supplies access to Microsoft 365 data. Semantic indexing improves discovery. Copilot connectors extend the available information. Work IQ assembles the context and tools into an agent-ready layer. Agent 365 adds centralized governance and observability around the agents using it.

This creates an important shift in enterprise AI architecture.

Organizations no longer need to treat every agent as an isolated application with its own Microsoft 365 retrieval system, permissions logic, indexing pipeline, and workplace integrations. Work IQ provides a reusable context layer that can support agents across applications, roles, and business processes.

However, Work IQ will also expose the strengths and weaknesses of the Microsoft 365 environment beneath it.

Organizations with accurate permissions, governed content, clear ownership, strong identity data, and thoughtful agent policies will be better positioned to create reliable agent experiences. Organizations with overshared data, duplicate documents, and unclear processes may find that AI makes those problems more visible.

The opportunity is not simply to help agents search Microsoft 365. It is to give them enough governed context to understand how work moves through the organization and help people move that work forward.

Emergent Software helps organizations plan, govern, build, and deploy Microsoft Copilot and custom AI agent solutions. Our team connects Microsoft 365, Copilot Studio, Microsoft Foundry, Fabric, security, and custom applications to create agents that work within the organization’s actual processes and data environment.

Explore Emergent Software’s Copilot and custom AI agent development services or contact our team to discuss how Work IQ could support your enterprise AI strategy.

Frequently Asked Questions

What is Microsoft Work IQ?

Work IQ is Microsoft’s workplace intelligence layer for Copilot and AI agents. It combines data from emails, meetings, documents, chats, people, and connected systems with organizational relationships, work patterns, tools, and permissions so agents can understand and act within a user’s workplace context.

Is Work IQ the same as Microsoft 365 Copilot?

No. Microsoft 365 Copilot is an AI experience available across Microsoft 365 applications. Work IQ is the underlying intelligence layer that helps Copilot and other agents understand workplace data, context, relationships, and available tools.

Is Work IQ the same as Microsoft Graph?

No. Microsoft Graph provides APIs for accessing Microsoft 365 resources such as email, files, users, calendars, and Teams. Work IQ builds on Graph and other signals to assemble relevant, agent-ready context and support reasoning and action.

What information can Work IQ access?

Depending on the agent, configuration, permissions, and product surface, Work IQ can use information from Outlook mail, calendars, Microsoft Teams, OneDrive, SharePoint, SharePoint Lists, people profiles, organizational data, and Microsoft 365 Copilot Search. External sources can also be connected through Microsoft 365 Copilot connectors and other supported integrations.

Does Work IQ give agents access to all company data?

No. Work IQ requests operate within the context of the signed-in user and applicable Microsoft 365 permissions. An agent does not automatically gain access to content that the user cannot access. Administrators must still review oversharing and ensure existing permissions reflect the organization’s intent.

Can Work IQ take actions in Microsoft 365?

Yes, depending on the available tools, integration surface, permissions, and agent configuration. Work IQ tools can support operations involving Microsoft 365 resources. Organizations should scope these tools carefully and require human approval for sensitive or high-impact actions.

What is Work IQ MCP?

Work IQ MCP is Microsoft’s Model Context Protocol interface for connecting agents to Microsoft 365 data and tools. Microsoft designed the interface around a compact set of generic tools that allow agents to retrieve information and act on Microsoft 365 resources through a simplified surface.

What are Work IQ workspaces?

Workspaces provide persistent working storage for longer-running agent tasks. Using SharePoint Embedded storage within the Microsoft 365 tenant boundary, agents can preserve intermediate outputs, continue multistep work, reuse results, and hand tasks between agents or experiences.

How is Work IQ different from a Copilot connector?

A Copilot connector makes external information available to Microsoft 365 Copilot and other intelligent experiences. Work IQ uses Microsoft 365 data, connected external information, relationships, work patterns, tools, and permissions to create a broader workplace intelligence layer for agents.

How does Work IQ relate to Fabric IQ?

Work IQ provides context about people, communication, documents, meetings, and how work gets done. Fabric IQ provides context about business data, metrics, entities, and operational relationships. An agent can use both to connect workplace activity with governed business information.

Is Work IQ generally available?

Microsoft made Work IQ API endpoints generally available on June 16, 2026. However, specific Work IQ capabilities and integrations may have different availability or preview statuses across Copilot Studio, Microsoft Foundry, Agent 365, and other product experiences. Organizations should review the latest Microsoft documentation for their intended implementation.

Where should an organization start with Work IQ?

Start with one clearly defined task that requires context across multiple Microsoft 365 sources, such as meeting preparation, a project handoff, or a customer briefing. Review permissions and content quality, connect only the necessary sources and tools, test with real users, and measure whether the agent improves the business process.