The Help Desk Job Is Shrinking. The Systems Admin Job Is Splitting Into Two.

September 17, 2026

Jillian, CMO— AI-assisted and reviewed prior to publication.

Computer support and general systems administration jobs are shrinking, while network architecture, systems analysis, and AI administration roles in the same IT department are growing. The dividing line is not whether you work in IT. It is whether your job is repetitive resolution work or judgment work that AI tools cannot yet do on their own.

Is AI Actually Shrinking IT Support Jobs?

Yes, and the government's own numbers now say so plainly. The Bureau of Labor Statistics projects that overall employment of computer support specialists will decline 3 percent from 2025 to 2035, and it names the mechanism directly: organizations continuing to implement automated tools, including chatbots, for troubleshooting.

That is not a prediction buried in a vendor white paper. It is the federal government's official occupational outlook, and it extends beyond the help desk. The same BLS data set shows network and computer systems administrator employment declining 4 percent over the same decade, even as the agency notes there will still be roughly 13,400 openings a year from people transferring out or retiring.

If your job today is mostly first-contact resolution, password resets, basic ticket triage, or routine server upkeep, this is the data you should be planning around, not ignoring. The roles are not vanishing overnight. They are shrinking relative to a growing tech workforce, which is a slower and easier trend to miss until you are the one competing for fewer openings.

Which IT Roles Are Growing While Support Shrinks?

The roles built around designing systems, analyzing them, and managing AI-driven investment are growing at multiples of the overall job market. BLS projects computer network architect employment to grow 8 percent through 2035, explicitly tying part of that demand to companies upgrading infrastructure to leverage AI. Computer systems analysts show the same 8 percent growth figure, with the agency citing organizations expanding IT and AI as the driver.

Here is how the fork looks side by side using BLS's 2025-2035 projections:

That is the same broad IT department, split by task type. Hiring demand is following the pattern you would expect: CompTIA's State of the Tech Workforce 2026 report found more than 275,000 active U.S. job postings in January 2026 referenced a need for some level of AI skill, spanning dedicated AI roles and jobs that simply expect AI fluency layered on top of existing responsibilities.

What Is the New AI Administrator Role, Exactly?

It is the clearest evidence that employers are creating distinct roles for AI oversight rather than folding it quietly into existing job descriptions. Microsoft built a dedicated AI Administrator role into Entra ID, and its own role reference documentation lists specific duties assigned to it: managing AI-related enterprise services and Copilot agents from the admin center, approving and publishing line-of-business Copilot agents, and managing the full lifecycle of agent identities.

That is a named, scoped, auditable job function, not a vague expectation that everyone gets "AI-savvy." A hiring manager filling that role is not screening for someone who has played with a chatbot. They are screening for someone who understands identity governance, consent policies, and agent lifecycle management well enough to be trusted with a privileged role. Research from ServiceNow, in partnership with Pearson, estimates that 31 percent of U.S.-based sysadmin roles will be transformed by AI over the next five years, with the field also needing new roles created specifically to manage AI tooling rather than simply losing headcount to it.

Put those two findings together and the picture is consistent. The administrator role is not disappearing. It is being redefined around governance of the tools that used to be someone else's job to build.

How Do You Position Yourself for the Growth Side of the Fork?

Start by being honest about which side of the table your current responsibilities sit on. If your daily work is closer to the support specialist row, the practical move is to add the specific skills tied to the growing rows, not to hope the shrinking row stabilizes.

A few concrete steps make that shift real instead of aspirational:

• Learn the identity and access fundamentals that sit underneath every AI administration role, since agent governance is built on the same consent, role, and permission models as traditional identity management.
• Get hands-on with the actual admin console duties employers are now naming in job postings: agent lifecycle management, consent policy configuration, and Copilot settings oversight, not just general AI awareness.
• Document the shift on your resume and LinkedIn profile in role-specific terms. "Familiar with AI tools" reads as filler. "Configured Copilot agent consent policies and managed agent identity lifecycle" reads as a credential.

Forge University's CompTIA and Microsoft certification resources walk through how these role-specific skills map to actual exam objectives, which is useful if you are trying to figure out which credential actually matches the job posting language you keep seeing. For the AI administrator path specifically, the AB-900: Copilot and Agent Administration Fundamentals certification is built directly around the governance, licensing, and agent management duties that Microsoft assigns to the Entra AI Administrator role, which makes it a more direct match than a general AI literacy course.

None of this requires abandoning the skills you already have. A support background is still the fastest way to understand how an organization's systems actually behave under real user load, which is exactly the judgment that keeps administrator and analyst roles resistant to full automation. The certification is what turns that background into something a hiring manager can verify quickly, rather than something you have to explain in an interview and hope they believe you.

If you want a study plan built around this specific transition rather than a generic IT refresher, you can start training on the objectives tied to the roles that are actually growing. The BLS numbers are not a forecast you can wait out. They describe a shift that is already showing up in job postings, and the professionals who move early are the ones filling the new openings rather than competing for the shrinking ones.

The uncomfortable part of this data is that it rewards specificity over general effort. Being generally good at IT support no longer differentiates you the way it did five years ago, because the routine parts of that job are exactly what automated tools are built to absorb. Being specifically certified in the governance and administration skills tied to AI systems does differentiate you, because that is the part employers have decided still needs a person with a name and a credential attached to it.

Start training free at Forge University