SCCM’s Long, Slow Death: Why Intune is Dancing on its Grave

System Center Configuration Manager (SCCM) has served IT teams for years, but its slow death is undeniable as Microsoft pivots toward Intune and cloud-first solutions. The shift reflects the changing needs of modern IT environments, driven by mobility, remote work, and cloud-based services. Let’s dive deeper into why SCCM’s days are numbered and what you can expect as it fades into the background.

 

1. Legacy Infrastructure Doesn’t Cut It Anymore

SCCM was a powerhouse in its prime, built for a time when most IT operations were centralized and on-premises. Enterprises relied on SCCM’s complex infrastructure, including SQL databases, distribution points, and local servers to deploy updates and applications. Back then, it made perfect sense—but in today’s cloud-driven world, these requirements are cumbersome. IT teams now prefer lightweight solutions that don’t require large on-prem servers and heavy management infrastructure.

SCCM's reliance on physical infrastructure makes it difficult for IT teams to manage today’s distributed, hybrid workforce, where devices are scattered across multiple locations. Intune’s cloud-first model, on the other hand, allows IT admins to manage devices from anywhere, without the need for physical servers.

With remote work here to stay, relying on SCCM feels like dragging along an anchor in an era of lightweight speedboats. Microsoft’s focus is increasingly on Microsoft Endpoint Manager and Intune, leaving SCCM to manage its decline while businesses adopt cloud solutions.

 

 

2. SCCM’s Bloated Costs and Complexity

SCCM’s infrastructure is not only heavy but also costly. From maintaining on-premises servers to ensuring databases are up to date, SCCM requires significant financial and human resources to run effectively. IT admins need to constantly monitor infrastructure, manage backups, update software, and troubleshoot issues. In contrast, Intune’s cloud-native design allows organizations to offload infrastructure management to Microsoft, drastically reducing costs and administrative burden.

Intune’s approach is far simpler. Updates are managed automatically in the cloud, reducing manual maintenance tasks and freeing up time for IT teams to focus on more strategic projects. Companies no longer have to worry about distribution points, patch management, or managing complex environments, as all of this is taken care of through Intune’s cloud service.

The Forrester Total Economic Impact (TEI) study on Intune shows that organizations switching from SCCM to Intune report up to 60% reduction in infrastructure costs, making it clear that sticking with SCCM is not just outdated, but financially inefficient.

 

3. Sluggish Updates and Innovation

SCCM is still receiving updates, but it’s clear that these are mostly maintenance updates rather than innovations. Microsoft’s strategic investment is going into Intune and its integration with Azure, Microsoft 365, and Unified Endpoint Management (UEM) features. SCCM’s development is noticeably slower, and the latest features are almost exclusively appearing in Intune.

For example, SCCM’s Windows Update management requires manual configurations and often takes longer to execute, whereas Intune offers Windows Update for Business or Autopatch, a cloud-based update management tool that streamlines the entire process. Intune can push updates to remote devices without the need for cumbersome VPNs or relying on devices being connected to the corporate network.

Intune’s roadmap is full of new features aimed at streamlining remote work, improving security, and automating device provisioning—all areas where SCCM struggles to keep up.

 

4. Deprecation of Key SCCM Features

Microsoft’s gradual deprecation of SCCM features is perhaps the clearest indicator of SCCM’s future. Legacy Reporting, for instance, has been replaced by Intune’s cloud-based analytics which offer more powerful insights without the infrastructure burden. App deployment, once a strong point of SCCM, has also been overtaken by Intune’s intuitive and cross-platform management capabilities.

As Microsoft continues to retire these core SCCM features, organizations relying heavily on SCCM are finding themselves at a crossroads. The writing is on the wall—while SCCM may still function, it’s on life support. Intune’s app management, with its ability to handle Windows, macOS, iOS, and Android devices seamlessly, has made SCCM’s application deployment process look antiquated.

 

5. Co-Management: The Transition Bridge

For organizations not ready to cut ties with SCCM entirely, Microsoft offers a co-management model that allows businesses to use both SCCM and Intune during the transition period. With co-management, IT teams can move workloads such as compliance policies, device configuration, and security management to Intune while continuing to use SCCM for legacy tasks.

Co-management offers a softer landing for companies transitioning from SCCM to Intune, but make no mistake—it’s still a transition, not a long-term solution. As SCCM continues to lose features and Intune gains new capabilities, the balance will inevitably tip towards full migration.

 

6. The Future is Cloud-Based

Ultimately, SCCM’s demise is part of the broader trend of IT moving to the cloud. The advantages of cloud-based solutions like Intune are undeniable—reduced costs, simplified management, scalability, and the ability to handle the remote work environments that define the modern workplace. The long, slow death of SCCM is just another step in this ongoing shift.

Microsoft is clearly placing its bets on Intune and the Microsoft Endpoint Manager ecosystem, and businesses that are still clinging to SCCM are running out of time. Planning for the future means planning for a world where cloud-based management is the norm, and SCCM is nothing more than a chapter in IT history.


Windows Enrollment Just Got a Glow-Up: Here’s the Good, the Bad, and the Patchy

Starting October 2024, new Windows 11 devices (version 22H2+) will automatically download and install quality updates during the Out-of-Box Experience (OOBE) when enrolled via MDM like Intune. This ensures devices are secure before user login.

 

 

thumbnail image 5 captioned Screenshot of the new progress percentage of setup.

Advantages:

  • Better Security: Immediate patching reduces risks from out-of-date software.
  • Compliance: Ensures policies are met from day one.

Disadvantages:

  • Longer Enrollment: Updates during OOBE might delay device handoff.
  • Access Pass Adjustments: Temporary Access Passes may expire before setup is complete.
  • Updates whether you want them or not: No more choice in the process.

 

 

 

 

 

 

 

Final Thoughts:

While the enhanced security is a significant benefit, the extended setup time could become a challenge for organizations with tight provisioning schedules. Admins managing large fleets may need to plan for the additional time required for these updates. The Temporary Access Pass issue also means careful coordination will be needed, particularly in environments relying on quick deployments or where devices are being "White Gloved" on behalf of Assistive Technology users.

For businesses prioritizing immediate security and compliance, this update is a step in the right direction. However, organizations focused on speed or with existing streamlined processes may face slight disruptions. Balancing security with efficiency will be key. One thing is for sure, Microsoft is keen for you to be on the latest updates!


Autopilot 2.0: Microsoft’s Slick New Rollout for Gov & Enterprise!

Microsoft has dropped a fresh upgrade for Windows Autopilot, and it's a game changer, especially for IT admins in government and large enterprises. The new features streamline the onboarding process, improve error handling, and provide support for government clouds like GCC High and DoD. If you're dealing with large-scale deployments, this update just made things smoother and less of a headache.

Key Features You’ll Love:

  1. Improved Out-of-Box Experience (OOBE): You can now see real-time progress as devices are onboarded. This not only gives you peace of mind but also lets you track issues before users get their hands on the device. The new UI is cleaner and provides more insight into the deployment process, which helps IT teams pinpoint problems quickly.
  2. Stronger Error Handling: The update includes enhanced error resiliency, meaning that even if something goes wrong during the setup, the system can handle it better, reducing the number of failed deployments. This saves time and frustration, especially in large environments where downtime can be costly.
  3. Government Cloud Support: For those managing sensitive environments, Autopilot now has extended support for government clouds, including GCC High and DoD. This means that government institutions with high security requirements can now enjoy the benefits of Autopilot without compromising on compliance or control.

thumbnail image 1 captioned A new Windows Autopilot device preparation section is available under Enrollment > Windows. Admins will configure a single Device preparation policy to configure deployment and user experience.Why It Matters:

The latest Autopilot updates are tailored for large-scale deployments, making it easier to manage thousands of devices while ensuring they stay secure and up-to-date. With better visibility into the onboarding process and a stronger focus on error resilience, IT departments can spend less time troubleshooting and more time focusing on strategic tasks.

For government bodies and enterprises that require strict security and compliance, these features are a blessing. The ability to deploy devices into secure environments like GCC High and DoD clouds without extra steps means faster, more efficient rollouts.

What’s the Catch?

While the updates bring a lot of improvements, it’s worth noting that device setup may still take some time, especially with larger batches. But overall, these enhancements mean fewer bumps along the way, less downtime, and a more intuitive deployment process.

Final Thoughts:

This Autopilot update is all about saving time, reducing errors, and providing greater control for IT admins working with large or highly secure environments. If you’re an IT admin in government or enterprise, these changes are worth exploring. Not only will they improve how you manage devices, but they’ll also give you more confidence in the deployment process from start to finish.


For more details, check out the full announcement here.


Intune Autopilot – Part 2 – Moving in the right direction!

Cont.

So if you've read part one, then you're looking for a way to set up an Intune managed Autopilot device to be shared among multiple users.

The chances are that you're already using autopilot for your single-user devices. Which means that you already have an Azure AD Dynamic Group, probably using a dynamic query such as:

(device.devicePhysicalIDs -any (_ -contains "[ZTDId]"))

The problem with this approach is that it's going to lump all of our autopilot devices into one AAD Dynamic group, which is no good if we need multiple deployment profiles (User Driven + Self-Deploying profiles). So we need a way to separate these devices. Enter the OrderID field...

The OrderID field, also referred to as GroupTag will allow us to create different groups of Autopilot devices, so that we can deploy Single User laptops, Shared PCs, and Kiosk devices all with different profiles and policy sets. However, if we are using the ZTDId tag, the chances are that none of our existing computers has the order IT field populated.

 

Therefore we need to populate the OrderID value for all existing devices and change our Dynamic AD groups to match the OrderID, not all autopilot devices.

Thankfully, Nicola Suter has come to the rescue here. He has written a PowerShell script that will connect to Graph API, and allow us to change the OrderID field in bulk. https://tech.nicolonsky.ch/bulk-update-windows-autopilot/ In this example. I'm going to use a group tag of "autopilot" to signify a standard user-driven deployed device, therefore we will use Nicola's scripts to change the group tag to "Autopilot" on all existing devices.

 

In conjunction, with this, we also need to change the Dynamic group query in Azure AD from targeting ZTDid to something like this:

(device.devicePhysicalIds -any _ -eq “[OrderID]:Value”)

Where "Value" is the value, you want to put in the OrderID field.

Once we have done this, we can start building a new policy and profile set for our other devices. In part 3, we will explore a multi-user desktop and kiosk mode deployment alongside traditional user-driven deployments.


Intune Autopilot - Sharing the love - Part 1

Over the last 18 months, I've had great success deploying modern-managed devices for clients, using a mix of Azure Active Directory, Intune and Autopilot. Autopilot has always worked great in scenarios where one user = one workstation, but what about a replacement for that traditional domain-joined PC which you would deploy for shift workers or in multi-user situations where one computer = 5 users?

There are a few ways that I have found a multi-user device can be set-up, but these all have drawbacks.

 

Build a PC then AAD join it:

On the face of it, this seems like the most straightforward option. The PC can be AAD joined and added into AAD Device groups to receive policies. However, the drawback with this approach is that the machine will be listed in AAD as a device under the user that AAD joined it. This can lead to admin accounts with tens of devices listed under them and a potential requirement to lift the "per user device limit" within AAD.

It's also a royal pain in the backside to have to assign group memberships manually, and policy application appears to be a little spotty the moment you transition to a standard user.

 

SharedPC CSP

Microsoft has thought of this and created a SharedPC CSP policy. SharedPC CSP comes with several advantages such as Guest account Management, Local Storage Policies, Power Policies, Fast first sign-in and Education Policies, all of which make management and sign-in much easier. But for me, this CSP has one critical flaw - it disables the OneDrive client completely!!

If like me, you've been encouraging people to take advantage included 1TB OneDrive space per user in Office 365, then you've probably migrated users' documents to OneDrive and are using policies such as KFM (known folder move). OneDrive client is critical to a Shared PC; otherwise, users will are forced to use OneDrive online to access their documents. It also gives no easy way to roam documents between shared PCs, for example at a reception desk with two seats.

 

 

So, how do we deploy an Intune Autopilot device for multiple users with all of our usual Line of Business apps?

 

The answer is Intune Self-Deploying Mode but without SharedPC CSP!!

 

What, I hear you cry, that sounds counter-intuitive, but it works. We'll let the self-deploying mode take care of our AAD join and then layer some custom Policies on top to cover the other bits.

The self-deploying mode requires adding another Autopilot deployment profile, which means you must make some changes to your existing autopilot configuration. In Part 2, I'll go into the changes that you need to make to prepare your environment for multiple deployment profiles.


Is ADFS now dead?

For those of you keeping an eager eye on cybersecurity, NCSE published some new guidance for securing Office 365 earlier this year. This new guidance includes one significant change from Microsoft, which some may find a little controversial.

Microsoft now recommends that hybrid environments – i.e. those that use Active Directory Domain Services and Azure AD – should prefer native authentication against Azure AD rather than ADFS.

In Microsoft-speak this is ‘Seamless SSO with Password Hash Sync’, configured to use either per-user or Conditional Access MFA.

Password synchronisation with the cloud can feel like a scary thing to do, but in actuality, organisations using Azure AD as their primary authentication source will lower their risk compared with ADFS.

This is because:

  • It’s the hashes of your password hashes that are sent to Azure AD, and not the reusable NTLM hashes commonly discussed in “pass the hash” attacks. This means that the credentials sent to Azure AD can’t be used to authenticate to any of your on-premise infrastructures that rely on Active Directory.
  • We are already relying on Azure AD to make access control decisions regulating who can see which data, hosted in Office 365. So we already need to trust that it’s built and operated securely. Storing password hashes doesn’t change that security requirement.
  • The availability of Office 365 will no longer be affected by any outages or downtime suffered by your on-premise ADFS or Active Directory infrastructure.

For those interested, the new Microsoft guidance can be found here:  It’s a brave new world out there!