If the current craze for artificial intelligence sounds familiar to you Peter CapelliBecause he’s seen the movie before, said Wharton management professor George W. Taylor. He noted that between 2015 and 2017, major consulting firms and the World Economic Forum confidently predicted that driverless trucks would Eliminate truck drivers Within a few years.
“You don’t have to think for long to realize that this doesn’t make sense in practice,” Cappelli told wealth exist skyrocketing from his home in Philadelphia.
“You don’t have to think very long about driverless trucks, okay, what happens when they need gas? You know? Or what happens if they have to stop to make a delivery? If they have to have an employee sitting next to them, of course that defeats the purpose, right?”
hat, who Recently partnered with Accenture in a Podcast Series To figure out the actual impact of AI on employment, we caution against listening too carefully to companies that are talking about their books or trying to sell you their new products.
“If you listen to technology developers, they’ll tell you what’s possible, but they’re not thinking about what’s practical.”
In the process of extensive conversations with everyone wealthCappelli tackles what artificial intelligence is really doing at work, just like him talked wealth Before The truth is, working remotely is terrible for most organizations.
“I mean, people say I’m a contrarian,” Cappelli said, “but I don’t think so, because I’m just skeptical of some things, you know?”
When pointed out that this was an inherently contrarian stance, Cappelli laughed and then got back to the point. “I’m just nervous about the hype.”
he and wealth On how his research fits into the broader picture of the second half of 2025, the second half of 2025 Influential MIT study attracts 95% of people’s attention Generative AI pilots failed to generate any meaningful returns. His favorite example is a specific case study of a company that actually put AI to work, both reducing headcount and increasing productivity. It still doesn’t quite match the predictions (e.g. from Elon Musk or Dario Amodei of Anthropicthis job will soon become optional or even a hobby). “The cost of doing this is very high,” Cappelli said of his discovery. “It was a success.”
three times the cost
Cappelli detailed the results of a case study he was involved in, Published in Harvard Business Reviewon insurance claims processor Ricoh: The exact type of low-level administrative tasks that AI should make easy to automate. However, the reality of adoption is a financial hit. While the company ultimately tripled its performance, the transformation was far from cheap. The company spent a year assembling a team of six, three of whom were expensive outside consultants, just to get the system running.
“The first thing they found,” Cappelli said, “was that large language models could do this very well for three times the cost of an employee doing it (manually). OK, so that didn’t work.” Costs included Ricoh paying an outside consultant about $500,000, Cappelli noted.
Even after optimizing the process, Ricoh still spends about $200,000 per month on AI fees—more than their total payroll for the task. He added that they cut headcount from 44 to 39 people, which shows how far away AI is from becoming a huge job killer in practice. His explanation was reminiscent of his self-driving truck example.
“The reason they still need employees is there are a lot of problems that need to be solved that are much harder to solve without AI,” he said. The good news, he added, is that the Ricoh unit’s productivity will eventually triple.
“That’s the reward, but it’s not cheap and it takes a long time to do.”
Ashok ShenoyRicoh America Vice President told wealth After AI began to be used to perform “very routine, repetitive, high-volume tasks,” human jobs did not disappear, but “shifted to areas where human judgment and experience are most valuable.” A year or so into the case study, he notes that Ricoh has successfully applied AI to mid-level, repetitive, time-consuming tasks at scale, and expects to use AI agents to automate some or all of its workflows in the next 6 to 12 months, “using human-machine interaction to resolve missing or unclear information and ensure quality.”
Shenoy acknowledged the significant costs highlighted by Cappelli, but noted that the project broke even in less than a year, and at $200,000 per month, it was cheaper than the previous operating model. “Although it does not rely on significant layoffs, the transition to artificial intelligence is expected to reduce overall costs by 15%.” Regarding the number of employees, he said that “this work is not driven by cost or layoffs,” and the implementation of artificial intelligence requires the creation of new roles, the redesign of existing roles, and the reassignment of team members to higher-value work. There have been no further layoffs, he said, with headcount remaining largely stable as productivity improves and output grows. “The bigger change is how people spend their time. They are doing less repetitive work and more focused on resolving exceptions, maintaining quality and serving customers.”
AI performance shame in the boardroom
Cappelli said he was working with Accentureit looks at MasterCardRoyal Bank of Scotland and Jabil. “These are success stories,” he said, and in the long run they will see productivity increases. Companies will be able to do more with fewer people, but “that’s going to take a long time to happen.” He thinks something crucial is being underestimated. “The key, though, is how much work it takes.”
Additionally, regarding layoffs, Cappelli said that at least in the areas he looked at, which are specific departments within each company, he didn’t see any. When Fortune contacted Accenture for comment, it said it generally agreed with Cappelli’s conclusions and cited CEO Julie Sweet’s sentiments. recent interviews wealth Editor-in-Chief Alison Shortell.
Cappelli believes that much of the noise surrounding artificial intelligence and the distance between possibility and practicality is driven by what other commentators have said “AI shame“.
Cappelli isn’t familiar with the term “AI shame,” but told wealth Describing what he saw was “absolutely correct”. “They’re pretending to say they’re doing something, right?” he said. “So there’s a lot of pressure on them to try to make these things work because investors like the idea.”
Professor quoted The Harris Poll found in early 2025 that 74% of CEOs People globally believe they will lose their jobs within two years if they cannot prove the success of AI, and about a third say they are actively adopting AI without truly understanding what it will bring. As the Harris Poll puts it: “CEOs estimate that more than a third (35%) of AI initiatives are little more than ‘AI cleansing’ for optics and reputation, but provide little if any real business value.”
Cappelli described how markets typically celebrate news of layoffs, even cited studies The “virtual layoffs” announced by companies never actually happen because companies are arbitraging the positive stock market reaction to news of potential layoffs.
Cappelli predicts there will be a “slow learning curve” and CFOs will start to realize “this is a very expensive thing.” The problem, Cappelli believes, is that American management has become “spoiled” and increasingly averse to the hard work of organizational change.
“(Employers) think it should be free. It should be cheap. You should be able to hang the shingles out and the right people will show up,” he said. In his view, true AI success requires “old-school HR” work: mapping workflows, breaking work down into tasks, and having employees work alongside AI “agents” to refine prompts.
“You can’t get ahead of your employees because employees do know how their jobs are done,” Cappelli said. The professor is frustrated with what he sees happening with most executives, saying they are essentially “avoiding” the real problems of solving the technology.
“They don’t see this as an organizational change issue and a big one,” he said. “They just put pressure on everybody, you know, to hope that somehow things will work themselves out.”