I'm Matt Lasker, the new community manager here at SuperOps. Part of my job this fall has been planning our IT Leader Roadshow, and that meant sitting down and actually reading the research on how AI is changing this job, not the vendor pitch version, the survey version, the one that asks people who do this work every day what's really going on.
What I found didn't match the demos I've sat through. Here's what six different surveys say, and what I think it means if you're the one holding the pager.
Your workload didn't get lighter. It got heavier.
Every vendor pitch promises AI clears your plate. The data says otherwise for most people. A 2026 survey of 844 IT professionals found that 52% report their workload has actually increased since adopting AI, and another 19% say it's stayed about the same. Barely anyone reports it getting lighter.
That's not because the tools don't work. The same survey found real time savings, 2.7 to 3.3 hours a week on specific tasks like triage and documentation. But faster tickets and a lighter day turned out to be two different things. The time AI saves on one task tends to get absorbed somewhere else, not handed back to you.
Source: SolarWinds, 2026 State of ITSM Report (844 IT professionals)
The cleanup tax is real, and it has a name now.
Vendor demos never show you the part where you're fixing what the AI just did. A survey of more than 1,600 full-time U.S. workers found people spend about 90 minutes a day with AI tools, roughly 47 days a year. Nearly half of that time, about 20 working days annually, goes to fixing its errors or rewriting prompts instead of doing the actual work.
It's common enough now to have its own name: workslop.
The time AI saves is real. So is the time it costs you cleaning up after it. From what I'm seeing in this data, the second number is quietly winning for a lot of admins.
Source: BambooHR survey, via HR Dive, covered by Metaintro
Fear is running way ahead of what's actually happening.
Ask around and you'll find plenty of anxiety about AI coming for IT jobs. I get it, I've heard it in plenty of conversations while planning this roadshow. A 2026 global survey of nearly 12,000 workers found 40% fear AI will make their job obsolete, up from 28% just two years earlier.
But when a sociologist went and actually asked 1,250 U.S. workers what had happened to them personally since 2023, the numbers looked nothing like the fear. Only 3% said they'd lost a job to AI. Meanwhile 6% landed a new AI-related role, and 9% got an AI-related promotion.
Four years after ChatGPT showed up, AI hasn't produced the mass displacement either the optimists or the pessimists predicted. It has produced a lot of anxiety that isn't matched by outcomes yet, which is its own kind of problem for anyone managing a team through it.
Sources: Mercer, Global Talent Trends 2026 (nearly 12,000 workers) and Jeffrey Dixon / YouGov, via The Conversation and Fortune (1,250 U.S. workers)
Nobody actually trusts AI to hit execute on its own. That's not a bug.
There's a version of this conversation, usually the one leadership hears, where admins are handing off decisions to AI wholesale. The data I found says otherwise. A 2026 survey on AI in IT found just 16% of organizations say they fully trust AI to make and execute decisions on its own. Another 36% say AI recommends but a human still decides everything, and 22% let AI act autonomously only in narrow, specific scenarios.
The trend is moving, just not toward full autonomy. 62% of respondents say their own trust in AI has grown over the last year, against just 5% who say it's dropped. Trust is going up. Full autonomy still isn't the norm anywhere. That tracks with a workforce that keeps double-checking its work, and for good reason.
Source: ITSM.tools / Atomicwork, 2026 State of AI in IT Report
Everyone says they're investing in AI. Half the workforce is training itself anyway.
A late-2025 survey of more than 27,000 digital workers and 1,225 employers found that roughly two-thirds of employers invested in AI tools over the past year. But only about half of tech professionals said their employer actually prepared them to use it. The other half are funding their own courses, docs, and YouTube time on their own clock.
If that's you, you're not behind. You're the majority. Honestly, this was the stat that stuck with me most while I was building this roadshow, because it's the one nobody puts on a slide.
Source: Randstad Digital, late-2025 survey (27,000+ digital workers, 1,225 employers)
The quieter shift: who you ask when you're stuck.
The last finding isn't really about the technology at all. It's about where knowledge lives on a team now. The same BambooHR-sourced survey found that a third of workers, 33%, say organizational knowledge transfer now happens primarily through AI tools instead of through colleagues. And 27% said they'd rather ask an AI a question than admit to a coworker that they needed help.
That one gave me pause. It's not a tools problem. If knowledge is quietly routing around the people who used to hold it, that's a team and culture problem, and it's one most AI rollouts aren't measuring at all.
Source: BambooHR survey, via HR Dive, covered by Metaintro
What I take away from all this
Put together, these six findings tell a pretty consistent story. AI is saving real, measurable time on specific tasks. It is not, for most IT admins, adding up to a lighter workload, because the saved hours are being reinvested in cleanup, verification, and work that didn't exist before. Fear about job security is outpacing the actual evidence of displacement. Trust is rising, slowly, but nobody serious is handing over the keys yet. And training is happening, just mostly on admins' own time rather than their employer's.
None of that is the story I've heard in a vendor demo. It's a lot closer to what I suspect is actually true in your building right now, and it's exactly why I wanted to build this roadshow around real data instead of another slide deck of predictions.
If you want to keep comparing notes on this, that's what the IT Admin Group is for, other people doing this work talking about what's actually changing, what's worth learning next, and who to actually ask when something breaks. No pitch, just the room.
Would love to hear what your experience has been. Let me know below in the comments!