A few years ago, employees were sneaking ChatGPT and Claude past corporate IT departments, setting up private accounts, and completing work that used to take them hours in a fraction of the time. An MIT study found that in more than 90% of companies, people used chatbots on a daily basis, even though only 40% of those companies had official subscriptions. Corporate leadership called it a security problem. Employees saw it as making their jobs easier. But now the situation has reversed completely.
A new global survey by WalkMe polled 3,750 managers and employees across 14 countries, and the results are surprising, to say the least. More than 54% of workers simply bypassed their company’s artificial intelligence tools in the past 30 days and did the work manually. Another 33% did not use AI at all. Taken together, this means that roughly 8 in 10 employees are either actively rejecting or avoiding the technology into which their companies are pouring record sums of money.
Average digital transformation budgets rose by 38% year over year to $54 million. Yet 40% of these investments are delivering below-expected results precisely because people simply do not use the technology. WalkMe CEO Dan Adika regularly meets with chief technology officers at large companies and asks them a simple question: how many of your people actually use AI for meaningful work? "The numbers are below ten percent," he adds.
Managers and employees live in different worlds
Corporate leaders and their employees appear to be describing completely different scenarios. Just 9% of employees believe AI can handle complex, business-critical decisions. Among managers? 61%. That is a gap of 52 percentage points. 88% of executives believe their people have adequate tools. Only 21% of employees agree. A 67-percentage-point gap over whether they even have anything they can use.
"Artificial intelligence has failed to deliver on its promises. Welcome to the real world. Forget the AI bubble. It simply has not delivered. Productivity? Weak. If AI really worked, productivity would rise significantly. But those Silicon Valley curves simply are not materializing." Economist Steve Hanke of Johns Hopkins University said on the subject.
Dan Adika has an apt metaphor for the entire situation. He says that buying every employee an artificial intelligence tool is like giving everyone a Ferrari. But these people do not know how to drive. They have no gasoline, meaning context and data. They do not know how to handle the car, meaning how to provide the right instructions. And at many companies, there are not even any roads, meaning APIs or other technical infrastructure, for them to drive on. What do you do when you have a Ferrari but no driver, no gasoline, and no roads? You will not get very far very quickly.
This gap costs companies almost two months of work each year
The losses can be quantified. The WalkMe survey found that employees lose 51 working days a year due to technological friction—that is, problems with tools that simply do not work as they should. That is nearly two months of lost time and an increase of 42% compared with 2025. It amounts to 7.9 hours per week. At the same time, Goldman Sachs calculated that those who use AI correctly save an average of 40 to 60 minutes a day. What AI gives to those who know how to use it, it takes away from those who struggle with it.
And then there is shadow AI. 78% of managers want to penalize the use of unapproved tools. Yet only 21% of employees say that anyone has even informed them about company policies. And 34% of workers do not even know which tools their company has actually approved. Meanwhile, 62% of these same managers privately admit that the risk of unapproved AI is overstated compared with the risk of companies not using AI at all.
Fear of being replaced?
Adika speaks of a new layer that has emerged beneath shadow AI. This is not about employees who break the rules. It is about those who do nothing at all. "They take pride in their work," he says of those who resist AI. "They will not let some robot replace them, and they will always find and point out the mistakes the tool makes." Sound familiar? That is exactly what the phenomenon of quiet quitting looked like, when people stopped doing anything extra during the COVID pandemic without formally leaving their jobs.
Oracle announced layoffs affecting tens of thousands of people. Block did the same. Critics call it "AI washing"—concealing previous overhiring behind a convenient excuse that also happens to please the market. Employees are not missing the implications of this logic. Adika says their fear is justified, but their conclusion is mistaken. "No bank or insurance company CEO is going to start laying off masses of people tomorrow. Who would do the work then?"
Three types of people. Racers, creators, and builders
Brad Brown of KPMG is thinking more than ever about how to bridge this gap. At the company, he began dividing employees into three categories: builders, creators, and experienced users. Each category is assigned specific career paths and incentives. "Now is the time for people to catch up with technology," he says.
The key idea behind this approach is that the problem is not people’s intelligence or conventional training. It is critical thinking and judgment that transform an ordinary employee into someone who truly uses AI. Not a technician. Not a programmer. One-third of corporate employees have never used any AI tool. These people report the lowest level of support from their employer, the least training, and the greatest concerns about the future. They are not rejecting AI; they simply do not know how to use it.
Hanke, who initially banned his students from using AI altogether, described it as a journey from "no" through "maybe" to "this is great, but some tools are garbage." AI now serves as his research assistant. "It saves me a ton of time. What would take an assistant a week in the library, I can do in an hour. But you have to know what AI is good for. And you have to understand the subject well enough to recognize the mistakes in the first place."
Additional source: aol.com
