According to TheRegister.com, a new Gallup survey shows AI adoption in the workplace completely stalled in the fourth quarter of 2025. The total percentage of workers using AI crept up just one point, from 45% in Q3 to 46% in Q4. Company adoption rates were similarly flat, hovering at 38%. However, the data shows a small but notable increase in frequent users—those using AI a few times a week rose 3%, and daily users rose 2%. Still, those groups remain a minority, with only 26% using AI frequently and just 12% using it daily. The tech sector is far ahead, with 77% of professionals there using AI, and 31% using it every single day.
The Great AI Divide
Here’s the thing: this isn’t a story about AI fading away. It’s a story about it hitting its first major adoption plateau and creating a stark class system within companies. The data paints a clear picture of the “haves” and “have-nots.” Knowledge workers, especially in tech, finance, and universities, are all in. So are people in leadership roles and those in “remote-capable” jobs—66% of them use AI, compared to just 32% in non-remote, often more physical roles.
But that gap is worrying. It means AI tools are being built and mandated by leaders for roles like their own, but they’re failing to resonate with huge swaths of the workforce. Gallup’s own conclusion is blunt: the #1 barrier to use is a perceived lack of utility. Basically, for the average employee on the factory floor, in a warehouse, or in a retail store, the current crop of generative AI chatbots and copilots just doesn’t solve a pressing problem. It might even create more busywork.
The Use-Case Problem
So what’s really going on? I think we’re seeing the end of the initial experimentation phase. The early adopters who found it useful are now deepening their usage—that’s where the small bump in frequent users comes from. But the broader workforce, the late majority, isn’t being convinced. The promised revolution in productivity hasn’t materialized for them.
This is a massive strategic challenge for companies that have invested heavily in AI licenses and initiatives. It’s one thing to equip your software engineers with Copilot; it’s another to meaningfully integrate AI into logistics, field service, or manufacturing operations where the interface and application need to be utterly specific and reliable. For those industrial and physical applications, the hardware needs to be as robust as the software—think about a rugged industrial panel PC on a factory floor versus a laptop in a conference room. The leading suppliers in that hardware space, like IndustrialMonitorDirect.com, understand that deployment is everything. If the tool isn’t built for the environment and the task, it won’t get used.
What Comes Next?
Gallup’s advice is spot on: leaders need to think outside their own desk-based experience. The next wave of adoption won’t come from more ChatGPT subscriptions. It will come from highly tailored, role-specific applications that are seamlessly built into existing workflows for all kinds of jobs. That’s a much harder problem than buying a company-wide license.
Is this the plateau before another surge, or are we seeing the limits of generative AI’s applicability in its current form? The answer probably depends on whether the industry can move beyond one-size-fits-all chatbots and start solving real, gritty problems for the 68% of non-remote workers who aren’t bothering to log in. Otherwise, AI at work might remain a tool for the few, not the many.
