Micha Kaufman, CEO of freelance marketplace Fiverr, had some hard truths to share with employees in a candid company-wide memo last year.
“AI is coming to your job. Look, AI is coming to my job. This is a wake-up call,” he warned. One year later, he has a message for executives looking to navigate the AI tsunami.
“Don’t be a cheerleader. If you’re not practicing, don’t preach,” Kaufman says. luck. “You can’t put AI on a wall and not act on it.” CEOs are currently treating AI as a training issue, buying products, holding seminars and checking boxes, but the real challenge is a cultural challenge that starts at the top, he said.
Across industries, there is palpable anxiety about the impending onslaught of AI and how best to prepare employees, managers, and most importantly, ourselves for the new reality that lies ahead. Technology is advancing faster than an organization’s strategy can keep up, and executives tasked with leading the transition often know this in real time. Additionally, many companies feel there is a gap between their AI ambitions and their outcomes. There is a lot of pilot work and hype, but only a few organizations, usually in the technology industry, are seeing transformative results.
Eight out of ten Fortune 500 companies are now using some kind of AI agent in their business built with “low-code/no-code/or vibecoding” tools. Agents handle everything from sales to customer service.
Source: Microsoft Research
“Many companies are struggling with some kind of dissonance between their expectations of AI and the reality of what they expected,” said Kate Smaje, senior partner and AI lead at McKinsey. “There are businesses all over the map.”
This disconnection keeps CEOs up at night. According to a recent study by Harris Poll, 79% of US CEOs believe they could lose their jobs within two years if they fail to achieve measurable business benefits from AI. Part of this is investor pressure for ROI, and part is FOMO. While some sectors, such as software engineering, are seeing significant productivity gains from AI, others are still struggling with how to implement basic tools.
It’s a daunting moment for business leaders trying to prepare. But with so much at stake, it is not an issue that can be ignored.
stick danger
One response to this fear was to move from an era of experimentation, where employees were encouraged to experiment with AI, to an era of top-down mandates and formal pilots, where employees were required to try out specific tools and demonstrate measurable results.
Companies like Meta, Amazon, Salesforce, and Microsoft are cracking down to force the adoption of AI within their workforces and mandate, monitor, and evaluate the use of AI tools. At Meta, a new performance appraisal system will allow engineers to track the number of lines of code they write with AI assistance, while Amazon managers will have a dashboard to monitor the use of individual AI tools to factor into promotion decisions, according to media reports.
What we want from a CEO is to think more like a scientist than an average person.
But technology companies have a history of driving workforce trends ahead of the rest of the business world. And of course, tech companies have a unique aspect to the game, as they manufacture and sell a variety of AI products.
Outside of technology, CEOs are less clear-cut. As Wharton management professor Peter Cappelli points out, too many executives are still “listening to the people who built the tools” rather than asking whether the same approach makes sense for their business. He argues that although builders are “not experts in business or management,” their success stories are treated as a universal blueprint.
Some companies are betting on peer-led learning and active incentives to accelerate adoption of AI tools, rather than mandates.
“If we take a solid approach now, I think some people may be able to achieve basically the right short-term goals, but not the long-term goals of building more agile and resilient organizations,” said Greg Hart, CEO of online learning platform Coursera. The stakes for companies successfully adapting to AI are higher than immediate productivity metrics. And with many employees viewing AI as a threat to their livelihoods, the mandate is likely to deepen rather than alleviate that anxiety.
According to recruitment firm Challenger, Gray & Christmas, companies will attribute about 55,000 job cuts directly to AI in 2025, more than three times the total for the past two years. When enterprise software company Atlassian cut 10% of its workforce in March and fintech company Bloc cut 40%, it did little to ease employee concerns. Block CEO Jack Dorsey said AI tools and “smaller, flatter teams” are fundamentally changing the nature of work and “what it means to start and run a company.” Some employees are also concerned that by using AI in the workplace, they are training an automated machine to replace it.
26%
Of the 2,300 companies surveyed, the percentage of companies that currently have a chief AI officer is up from 11% in 2022. More than half were appointed from within the company.
Source: IBM Business Value Institute
Fiverr’s Kaufman argues that this is precisely why leaders need to separate fear of AI from AI skills. He says companies often “collapse” the separation anxiety conversation and the upskilling conversation, making both worse in the process. Kaufman said concerns about evictions are “legitimate” and deserve to be discussed directly and honestly, not as a “corporate security play.” Only once it’s on the table can leaders talk with confidence about how roles will change, what categories of work will shrink or grow, and what new skills people will actually need to develop.
A CEO who is a scientist, not a general
Joseph B. Fuller, professor of management practice at Harvard Business School, says companies “have to get used to” spending more time learning now and resisting pressure to take premature actions that they may later regret. What’s needed is a CEO who thinks more like a scientist than a layman, someone who is comfortable not only overseeing experiments but also protecting those running them from penalties when things don’t go as planned.
The job of a successful CEO is to “create the conditions for risk-free experimentation by making sure that the people conducting the experiment understand that their senior and best colleagues understand:” [and] Rather than quietly shelving AI pilot projects that don’t yield results, Fuller recommends celebrating well-run failures and sharing knowledge.
Coursera’s Hart emphasizes the importance of using this early stage of the AI era to learn and adjust.
“If we only focus on efficiency right now, we are missing out on the opportunity to think about the true transformational impact AI can have on companies, given that we are still in the early stages of what it can do,” he says.
Coursera hosts monthly “AI Spark Sessions” where employees volunteer to share how they are leveraging AI to make their jobs easier and more effective. Hart said these sessions are the most well-attended across the company, with staff openly sharing tools, workflows and follow-up resources without hiding the efficiencies they’ve discovered.
“If you take a solid approach, some people may get their short-term goals right, but not their long-term goals.”
Greg Hart, Coursera CEO
This is especially important for AI projects where the return on investment is not always immediate. Economists call this the “J-curve.” Because companies absorb learning costs before making profits, productivity declines before it soars.
Last year, when a now infamous MIT report found that the vast majority of AI pilots do not make meaningful profits, investors panicked and treated this as an indictment of AI technology. In fact, the report found that the biggest cause of underperformance is not the technology itself, but a broader “learning gap” in which large organizations lack the expertise to meaningfully incorporate AI into their workflows. It turns out that startups that are freed from the burden of entrenched processes and office politics fare much better.
Think beyond technology
It helps to remember that executives have been here before, and there are valuable lessons from the past. When the Internet arrived in the 1990s, the last time the technology promised to reshape business, most companies embraced it and hoped for the best.
In the early dot-com era, companies tended to treat the Web like a digital brochure rack. It was treated more as a shiny distribution channel than a reason to rethink how the web works. It wasn’t until a small number of companies began to rebuild their businesses around the web that things really changed for others.
It wasn’t access to technology that separated the winners from the laggards. It was whether leaders were willing to challenge habits, redesign work, and tolerate periods of disruptive experimentation.
In that sense, AI may not be that different.
“If we just deploy AI, we’re already seeing evidence that AI isn’t going to live up to expectations,” said Aneesh Raman, chief economic opportunity officer at LinkedIn. “Providing employees with proficiency in how to use AI only gets them halfway. The real impact comes when workers use AI to transform their jobs – not just adding another tool, but redesigning tasks and workflows.”
This article appears in the April/May 2026 issue of the magazine. luck The headline: “Executives Adapt or Die.”
