How to Become a Desktop Engineer — Part 8: SCCM & Co-management

How to Become a Desktop Engineer — Part 8: SCCM & Co-management

How to Become a Desktop Engineer — Part 8 of 13

SCCM & Co-management

Post 7 gave you Intune — the cloud console, one policy reaching every device at once. This is the eighth post in a thirteen-part series on how to become a desktop engineer, and it's the honest other side: the half of most estates that still runs on Configuration Manager, why big shops still demand it, and how co-management moves workloads across.

The postings that list Intune also list SCCM: Simtra names both as the primary endpoint manager, and Horizon's senior posting demands expert-level SCCM. The cloud-first story and the on-prem reality aren't two different jobs. They're the same job, in the same estate, on the same day.

What SCCM still owns

Configuration Manager — SCCM to everyone who has worked with it, ConfigMgr to Microsoft — is the on-premises management platform, and it still owns the work Intune doesn't do as well.

OS deployment is the big one. A task sequence is a scripted build: partition the disk, lay down the image, install drivers, join the domain, install the base apps — hundreds of steps, in order. Autopilot hands a new machine to the user and builds it around them; a task sequence builds a machine on the bench, exactly the same way, five hundred times. For complex imaging — a lab machine with a hundred gigabytes of specialist software, a locked-down kiosk — it's still the tool.

Collections slice the fleet: queries grouping devices by OS version, model, site or installed apps — the on-prem equivalent of Post 7's dynamic groups: target the collection, never the device.

Software and patch deployment run through the same engine: applications pushed to collections with detection rules, and software updates — including Windows Server patching — from the same console. That last part matters: Intune manages Windows clients. ConfigMgr manages the servers too.

Behind it all: management points that talk to the clients, distribution points holding content close to where it's needed — the branch office pulls a ten-gigabyte image from its own server instead of across a WAN link — and the site server holding the database. When a deployment stalls, the first question is usually a content question: is the package on the distribution point the client is asking?

The honest state of play

Diagram of Microsoft's Configuration Manager annual release cadence announcement: version 2609 is the first annual release in September 2026, with 18 months of support per version, and three actions — keep ConfigMgr current, move workloads one at a time, learn both consoles.

Both sides of the story are true, and the engineer who can say both out loud is the one hiring managers trust.

Microsoft's side, in its own words: on November 5, 2025, the Configuration Manager team announced that starting with version 2609 — September 2026 — ConfigMgr moves to one release a year. "Microsoft Intune is the future of device management, and all new innovations will occur there. Configuration Manager will continue to serve your on-premises devices, with a renewed focus on security, stability, and long-term support." Each version gets 18 months of support; hotfix rollups arrive only when absolutely necessary.

No end-of-life has been announced — the rumour mill has announced one about once a year for a decade. What Microsoft actually said is narrower: new features go to Intune; ConfigMgr gets security, stability and long-term support. A maintenance future, not a funeral.

The other side: senior and enterprise roles in October and November 2026 still demand expert SCCM, because the estates those roles run are measured in tens of thousands of devices, with branch offices, imaging benches and servers that will never touch the cloud. A skill doesn't stop paying because the vendor's roadmap moved on — it stops paying when the estates stop needing it, and they haven't.

Learn it in that order: Intune first, because that's where the new work is; SCCM second, because that's where the big estates still live.

Co-management: the bridge, not the destination

Diagram of the co-management workload decision: four estate signals feed a decision — where does this workload live — with Intune and co-managed workloads reaching the cloud-managed fleet while ConfigMgr workloads stay on-premises.

Co-management is the answer to the question every estate eventually asks: we have ConfigMgr, we want Intune, and we can't switch overnight. Don't switch. Run both, on the same device, and move one workload at a time.

A device becomes co-managed when it is enrolled in Intune and still has the ConfigMgr client — it shows up in both consoles. Then, workload by workload, you decide which console is the authority for what. The workloads are the same levers from Post 7, split across two homes: compliance policies, device configuration, Windows Update policies, Endpoint Protection, Office Click-to-Run apps, client apps, resource access policies. Each one has a slider, and each slider moves independently. The order is a judgment call: compliance policies often move first — Intune's compliance engine is the one Conditional Access listens to — while client apps and OS deployment usually move last, because that's the work ConfigMgr does best and the riskiest to get wrong.

What "co-managed" looks like day to day: a device in both consoles, each workload with one clear owner, and the engineer's job is knowing which console to open for which problem. The failure mode is Post 7's group-assignment mistake, wearing a different coat: two consoles both thinking they own the same workload. One owner per workload. No exceptions.

The migration mindset

Moving an estate from SCCM-first to Intune-first is a migration, won the same way patch rings are won: pilot first, fleet second, rollback in your pocket.

Pick a pilot group that looks like the fleet — a few machines of each model, a few users in each department, one awkward edge case. Move one workload. Watch it for a full patch cycle, not a full afternoon. A compliance policy that looks fine on day one can block a VIP's device on day fourteen — and the pilot is where you want to learn that, not the fleet.

Rollback thinking is the discipline: before you move a workload, know exactly how you'd move it back. The slider goes both ways, and the engineer who has rehearsed the way back moves fast. The one who can't roll back moves slowly, or never at all — and an estate that never moves falls behind.

And the same rule from Post 3 applies at fleet scale: document the move — workload, date, pilot results, rollback steps — so the answer is in the notes six months later.

When SCCM is the right answer

There are estates where SCCM isn't the legacy to migrate away from — it's the right answer, and saying so is the expert move.

Large estates with complex imaging. A task sequence that builds a specialised machine identically, hundreds of times, is work Autopilot was never designed for: Autopilot builds a machine for a user; a task sequence builds a machine for a purpose.

Branch offices and content distribution. A distribution point in the branch means a ten-gigabyte image crosses the WAN once, then serves the whole office locally — instead of a hundred downloads of the same ten gigabytes.

Air-gapped estates. Some networks never touch the internet — defence, critical infrastructure, regulated labs. Intune is a cloud service; it can't manage what it can't reach. ConfigMgr can.

Servers. Intune doesn't manage Windows Server. ConfigMgr does — patching, inventory, software deployment, the lot. Any estate with a server room has a ConfigMgr argument.

Knowing when SCCM is the right answer — and explaining why in the language of the estate's constraints — is what separates the engineer who follows the roadmap from the engineer who serves the business.

One note if you learned imaging on MDT: Microsoft retired it in January 2026 and it doesn't support Windows 11. The skill transfers — a task sequence is the same thinking, with a supported engine under it — but the tool doesn't.

Practise it this week

  • Map a real estate. Take the lab from Post 7 — or a workplace you know — and write down which workloads live where: what would you keep in ConfigMgr, what would you move first, and why. The list is the thinking; the consoles come later.
  • Read the announcement. Microsoft's November 5, 2025 post on the annual release cadence is three minutes long. Read it in full, including the support table — then read the comments. The gap between what Microsoft said and what admins heard is the lesson of this post.
  • Move one workload in a lab. In a trial tenant with a co-managed VM, flip the compliance workload from ConfigMgr to Intune and watch the device appear in both consoles, then flip it back. The slider goes both ways — prove it before you touch a real fleet.

Next up: App Packaging & Patching. You've got the platforms; now the work that touches every user every week — turning software into packages that deploy cleanly, and keeping the fleet patched without breaking it.

Missed the earlier posts? Start with Part 1: What a Desktop Engineer Actually Does, Part 2: The Path, Stage by Stage, Part 3: The Mindset — Curiosity Is a Technical Skill, Part 4: Windows 11 Inside Out, Part 5: Identity: AD, GPO, Entra ID, Conditional Access, Part 6: Networking Essentials, and Part 7: Intune + Autopilot.

Comments

This site uses Akismet to reduce spam. Learn how your comment data is processed.