Employers are boosting productivity expectations even as workers struggle to translate AI confidence into results, according to a survey of 2,000 workers published Tuesday by WalkMe.
More than half of respondents said AI has led their manager or team to expect more output in the same amount of time, the adoption platform provider found.
From the top of the executive chain, business leaders are passing down the pressure to perform using AI. More than half of employees said senior leaders don’t fully understand their company’s public AI strategy.
“The higher you go in a company, the more they believe in AI,” WalkMe Global Field CTO KJ Kusch told Channel Dive. “But when you get to boots-on-the-ground, there’s a struggle with how to use it. There’s people who believe in it for context or writing, but I don’t think the real meat of ‘does AI do the tasks that matter the most?’ is really there yet.”
Some employees are developing their own strategies to navigate the AI productivity squeeze. One-third of respondents reported pretending to be more skilled in AI than they really are. Another third said they have passed off AI-generated work as their own.
Yet confidence in AI remains high. Nine in 10 respondents reported feeling confident with AI tools, though only a quarter said AI worked on their first try.
The survey points to a “confidence trap,” Kusch said. Employees say they’re AI-fluent but the results don’t match the claim.
“I’m not sure AI is giving results that really provide value to a company, like driving revenue or finishing business-critical work,” she said. “But employees are using it. They have better presentations than I’ve ever seen, but I’m not sure that actually matters. They write the best performance reviews, they send the best emails, but I think people wish it did more.”
The survey also found that there’s a generational gap in AI confidence. More than 9 in 10 Gen Z workers reported feeling confident in AI, though they were also the most likely to overstate their skills. Nearly half of the Gen Z cohort surveyed admitted they overstated their AI proficiency, compared to 13% of Baby Boomers.
“It’s not so much pretending as it is ‘you don’t know what you don’t know,’” Kusch said. “If an AI agent suggests something, you might now know if it’s wrong or right, so you just believe it. I’m not shocked that they’re more confident, because they’re not the ones typically finding errors. It’s the people who have been there, with more work experience.”
Kusch added that one of the most effective strategies for AI implementation she’s seen is using AI characters to act as on-screen, interactable coaches that help employees make decisions rather than instill fear about job security.
“It’s like bringing Microsoft Clippy back,” she said. “Instead of asking employees to put in answers, it asks them to help validate outputs, so the human feels like they’re in the loop of AI. Whereas I think the way AI is mostly delivered is through answers prompts, a bunch of work and an agent telling you whether you’re right or wrong at the end. It’s not a tool; it’s like filling in a ginormous comment box.”
