Microsoft Ignite 2026: 5 Announcements That Could Change the Modern Workspace

Microsoft Ignite 2026 comes with a lot of announcements. Instead of trying to cover everything, I picked the five developments that, in my opinion, matter most from a Modern Workspace perspective.
For years, we have mainly looked at the Workspace as: Users + Devices + Applications + Data
But AI agents are becoming part of that environment as well.
When agents start performing work for us, some familiar Workspace questions start to appear. Where do they run? What can they access? How do we manage them? And how do we make sure they meet our security and compliance requirements?
Looking at the announcements around Microsoft Copilot, Windows 365, Intune and Entra, I think we are starting to see how AI agents could become part of the Workspace we already manage today.
There is much more in the Book of News than I can cover here. For each of my five highlights, I've added the Book of News chapter and section, together with a link to the original Microsoft announcement so you can explore the details yourself.
You can find the complete Microsoft Ignite 2026 Book of News here.
Note: I wrote this on 20 November 2026, based on what Microsoft has shared so far. Some of these announcements are still in preview or have no release date yet, so details may change. I'll update the post when I learn more.
Let's take a look at my five highlights.
1. From Copilot to a Digital Colleague
We are getting used to Microsoft Copilot as an AI assistant. We ask a question, Copilot provides an answer and we continue working. With Autopilot, Microsoft is taking this a step further.
Autopilot, and yes, the name is confusing: it has nothing to do with Windows Autopilot in Intune. I would have picked a different name 😏. It is a persistent personal productivity agent with its own identity, memory and tools. Instead of waiting for the next prompt, the agent can continue working and perform recurring tasks.
For me, the concept of a digital colleague becomes more relatable here.
Today: User → Prompt → Copilot → Response Tomorrow: User → Delegates task → Agent continues working
And that raises some questions.
Who owns the agent? What information can it access? Which applications can it use? And how do we know whether an action was performed by the user or by the agent?
These aren't only AI questions anymore. They are Workspace questions.
Book of News
Chapter: Microsoft Copilot
Section: Autopilot is rolling out to Frontier customers
2. Windows 365 for Agents: A Workspace for Digital Colleagues
This is probably the announcement that caught my attention the most.
We already use Windows 365 to provide users with a managed Cloud PC. With Windows 365 for Agents, Microsoft applies a similar concept to AI agents.
Agents get a managed Windows environment in which they can perform computer-based tasks. Microsoft is also connecting Windows 365 for Agents with Copilot Studio and Microsoft Foundry.
But there is one scenario I find especially interesting. Agents will be able to access on-premises applications using their own agent identity.
Why does that matter?
Not every enterprise application has a modern API. Most organizations still have applications that were never designed with modern automation in mind. A computer-using agent running inside a controlled Cloud PC creates another way to interact with these applications.
That gives us a simple parallel:
Employee → Windows / Windows 365
Digital colleague → Windows 365 for Agents
This makes Windows 365 for Agents more than another AI service.
It becomes part of the Workspace architecture.
Book of News
Chapter: Windows
Section: Windows 365 for Agents scales to Microsoft Foundry, introduces persistent Cloud PCs
👉 Windows 365 for Agents in the Microsoft Ignite 2026 Book of News
3. Modern Management Is Expanding Beyond the Endpoint
Ignite isn't only about AI agents. There are also some notable changes for Microsoft Intune and Windows management. One that caught my attention is Declarative Device Management.
Instead of configuring every individual setting, IT defines the desired state and Windows works to apply and maintain that state.
That changes the approach from:
How should I configure the device? to What state should the device be in?
Also worth a closer look is Enterprise App Management, with Microsoft expanding automatic application updates beyond applications originally deployed by IT. For organizations managing third-party applications such as Chrome and 7-Zip, this could be very useful. An application doesn't suddenly stop being part of your attack surface because IT didn't originally install it. I have written several posts about EAM before (auto-update EAM apps and work-ready Autopilot devices), so when I get a chance to test this myself, I will certainly write about it.
And What About Windows Server?
There is another announcement that could easily be missed from a Workspace perspective.
With Windows Server 2028, Microsoft is adding another scenario to Microsoft Intune, alongside Azure Arc and Windows Admin Center.
Intune has continued to expand beyond traditional endpoint management, and Windows Server adds another piece to that journey.
It's still unclear where Intune will fit next to Azure Arc and Windows Admin Center, and I'll be watching how that develops.
Book of News
Chapter: Windows
Section: Windows and Microsoft Intune capabilities
👉 Windows and Microsoft Intune capabilities
Chapter: Azure
Section: Introducing Windows Server 2028
👉 Windows Server 2028 in the Microsoft Ignite 2026 Book of News
4. AI Agents Become Part of Device Compliance
This is probably one of the announcements I will be watching most closely from an Intune perspective and I'm looking forward to testing this new feature.
For years, we have used Intune Compliance together with Entra Conditional Access to determine whether a device is trusted enough to access company resources.
AI agents now add another factor to that decision. What happens when an unauthorized local AI agent is running on a corporate endpoint? Microsoft is bringing local AI agent governance into the endpoint management and security model. This creates the possibility of using the presence of unwanted agents as part of the compliance decision.
The model then starts to look more like:
Device + AI Agent
↓
Intune Compliance
↓
Entra Conditional Access
↓
Microsoft 365 & Corporate Data
I think this is quite an important change. Until now, device compliance was mainly about questions such as:
Is BitLocker enabled? Is the device healthy? Does it meet our security requirements?
Now another question is being added:
Does this device comply with our AI policy?
Book of News
Chapter: Security
Section: Intune
👉 Microsoft Intune helps secure AI adoption and reduce endpoint risk
5. Zero Trust Now Needs to Include the Agent
Compliance is one piece of the puzzle. Identity is another.
If an agent starts performing work independently, I want to be able to distinguish between an action performed by:
User → Application → Agent
This is where Microsoft Entra Agent ID comes in.
Instead of an agent simply operating underneath a user's identity, supported agents can get their own identity.
That allows us to start asking familiar Zero Trust questions:
- Who owns this agent?
- What can it access?
- Does it still need that access?
- What actions has it performed?
- Should we still trust it?
But having an identity doesn't automatically mean that an agent should always remain trusted. Microsoft is therefore also introducing Dynamic Agent Risk.
We are already familiar with:
User Risk + Device Risk
Now we need to start thinking about:
User Risk + Device Risk + Agent Risk
An agent might be approved today, but its permissions can change. Its behaviour can change. New capabilities can be added or the underlying model can change. That means agent security needs to become continuous. For me, this is a logical extension of Zero Trust:
Never trust, always verify now also applies to digital colleagues.
Book of News
Chapter: Security
Section: Microsoft Entra expands secure identity and access for agentic AI era
👉 Microsoft Entra Agent ID in the Microsoft Ignite 2026 Book of News
Chapter: Security
Section: Dynamic Agent Risk within Agent 365 in preview
👉 Dynamic Agent Risk in the Microsoft Ignite 2026 Book of News
Final Thoughts
Looking at these announcements separately, they may seem like individual product updates. But together, they show a real shift in the Modern Workspace.
Agents bring familiar Workspace questions with them. Where do they run? What can they access? How do we manage them? And how do we determine whether they can be trusted?
This is where technologies such as Intune, Entra, Windows 365 and Defender start to play a role beyond the traditional user and device.
The Modern Workspace isn't being replaced. It's simply expanding to include AI agents.
That is what I mean by a Secure AI-Powered Workspace: a managed and secure environment where people and agents can work with the applications and data they need.
