Research Roundup: AI on Org Charts, Job-Hoppers’ Skillsets, Autorenewal Subscriptions, and More

Staying on top of the latest research can help you make the most informed decisions for your company and employees. To help with this, we’ve pulled together useful and surprising insights from around academia covering topics including: Why you shouldn’t treat AI agents like employees Frequent job-hoppers’ hidden talent A better way to justify price increases For some pieces, we’ve also included links to relevant HBR.org articles so you can dig deeper into the findings. Why You Shouldn’t Treat AI Agents Like Employees As more organizations deploy AI, some aren’t just rolling out tools; they’re giving their agents names, roles, and performance goals. But this seemingly symbolic choice can change how managers oversee work and assign responsibility. In January 2026 researchers surveyed 1,261 HR and finance managers, directors, and executives across industries from the United States, Canada, and the European Union. About a third said that their leaders already talked about AI as a “teammate” or “employee,” and 23% reported that their organizations had listed AI agents on their org charts, assigning them formal roles. To see how that framing affects behavior, researchers ran an experiment in which 813 of those managers were asked to review documents—including job descriptions and budget reports—that contained built-in errors. They were given 20 minutes to review as many reports as they could, flag any errors, and escalate large problems to a higher-up. All managers saw the same documents. What varied was who they were told had drafted them: an AI tool, an AI employee listed as a direct report on an org chart, or a human employee who was a direct report. The framing had relatively small effects, except in one group: managers who already worked in firms that listed AI agents on their org charts. Among those participants, describing the drafter as an AI employee rather than an AI tool reduced their confidence in their own judgment. They caught about 18% fewer errors and were 22% more likely to escalate issues, even when they were told that they would be penalized if they asked for an unnecessary extra review. When asked who was responsible for the errors, these managers assigned about nine percentage points less responsibility to themselves and eight percentage points more to the AI if it was framed as an employee rather than a tool. They also shifted more blame to their team members and leadership. In organizations where AI agents are treated as teammates, managers are accustomed to doing less careful checking themselves. They also lean more on others to review work and are more inclined to see the AI as bearing responsibility if something goes wrong, say the researchers. “The evidence presented in our study suggests that AI agents should be treated as what they are—software automation,” which requires human accountability to minimize risks, they write. Go deeper: “Research: Why You Shouldn’t Treat AI Agents Like Employees,” by Matthew Kropp, Julie Bedard, Emma Wiles, Megan Hsu, and Lisa Krayer. Job Hoppers Have a Hidden Talent Hiring managers often treat frequent job changes as a red flag indicating someone won’t commit long enough to justify the onboarding investment. But a new study finds that people who switch employers a lot possess valuable social skills that help them get up to speed more quickly than other new hires. Using monthly employment and performance data on 8,693 U.S. hedge fund managers from 2004 to 2019, researchers tracked what happened when a fund manager joined a new firm. Consistent with past work, they found that 72% of moves were followed by a performance decline, from which it took an average of four months to recover. But that slump was not the same for everyone. Managers with more prior company moves suffered a smaller initial decline and returned to their premove performance within two months, compared with five months for those who had low mobility. The researchers suggest that repeated transitions create a form of transferable knowledge about how to get up to speed in new environments. With each move, people practice noticing unwritten norms, interpreting cues, and adjusting how they work with colleagues. Over time those skills become more generalizable and help make the next switch go more smoothly. The study found that the advantage was biggest when the social and cultural hurdles of joining were higher, such as when a manager joined a firm in a distant geographic location, and where existing employees had long tenures (and therefore norms are likely to be less flexible). The researchers note that more-mobile managers weren’t better performers overall; they were better at the transition itself and absorbed the shock of a move faster. Still, hiring teams may do well to reframe their thinking about “job hoppers,” from pure flight risks to something rarer and more useful: employees who can swiftly learn the ropes of a new organization. Go deeper: “The Case for Hiring Job Hoppers,” by Rebecca Kehoe and F. Scott Bentley. Giving Stock Options to Frontline Workers Improves Safety Offering stock options to your employees is usually framed as a way to build loyalty and align incentives with shareholders. When you extend those options to rank-and-file workers, you can also help make your workplace safer. Researchers matched injury data from the U.S. Occupational Safety and Health Administration with compensation data from 20,160 workplaces representing 629 firms from 2002 to 2011. Across the sample, which covered a wide range of industries—including retail, pharmaceutical products, transportation, and industrial metal mining—workplaces averaged about 6.6 injuries per 100 employees per year. After controlling for factors such as company size, use of seasonal workers, hours worked, and capital intensity, the researchers found that companies with higher-than-average stock options for regular employees had an injury rate that was 7% lower than that of companies offering average stock options. To examine causality, the researchers looked at what happened after a new federal accounting rule (FAS 123R) made stock options look more expensive on companies’ financial statements, prompting many firms to cut back on the options they granted. At the companies that had used rank-and-file options most heavily before the rule, injury rates rose more after its adoption than they did at otherwise similar firms, consistent with the idea that reducing options leads to worse safety outcomes. The researchers list two reasons for the connection between options and safety: retention and cooperation. Because unvested options tie employees’ wealth to the firm and are forfeited if employees leave, they can reduce turnover, and experienced workers are less likely to be injured than new hires. Options also tie rewards to team performance, which encourages coworkers to monitor and support one another on safety-critical tasks. For leaders, the takeaway is to treat broad-based stock options as more than a recruiting perk. When thoughtfully designed, they can reduce turnover, strengthen peer accountability, and improve safety, especially in high-churn, hard-to-monitor parts of the business where injuries are most likely. Is Your Autorenew Keeping the Right Customers? Many firms assume that making subscriptions “sticky” using trials that autorenew is an easy way to grow revenue. A new experiment shows that that model could be deterring engaged consumers from subscribing. Researchers partnered with a major European digital newspaper that erects a paywall after readers consume a quota of free articles. In 2018 more than 1.4 million readers who hit the paywall were randomly shown one of two offers: a trial that automatically rolled into a paid subscription unless canceled, or a trial that automatically ended unless the reader actively renewed. The researchers then tracked every reader’s subscription and site usage for roughly 20 months. In the short term, autorenewal did raise revenue. Among those who took the promotional offer, more remained paying subscribers in the months immediately after the trial than those in the autocancellation group. But the autorenewal offer sharply discouraged people from signing up in the first place. During the promotional window, 35% fewer readers subscribed when the deal was autorenewing rather than autocanceling. Over the course of the study, the initial gains made in the autorenewal group eroded and reversed, and the autocancellation offer ultimately produced 23% more total paid subscribers. Additionally, consumers who were offered autorenewal became less likely to subscribe to the publication through any channel afterward, including through future subscription offers. Autorenewal is so damaging because it is predicated on the assumption that consumers don’t realize they’ll forget to cancel a subscription they don’t use. But by examining usage data, the researchers found that consumers are actually self-aware. Readers who suspected they wouldn’t remember to cancel were put off by autorenewal and stayed away. Those who did sign up and found themselves paying for something they didn’t use later disengaged from the brand entirely. So should companies always use autocancellation? In an article on HBR.org that expanded on the findings, researchers Klaus M. Miller and Z. John Zhang note that the results best fit “stickier” products, such as news, music streaming services, or cloud storage—categories where customers usually stay once they find a good option. In markets where switching is more common—such as television streaming services, fashion subscriptions, or meal kits—autorenewal may play a more positive role. In industries with very high repurchase rates (above roughly 70%), the researchers say, autocancel may help lock in customers who want to be there for the long haul. Go deeper: “Should Your Subscription Business Use Auto-Renew?” by Klaus M. Miller and Z. John Zhang. To Justify Price Increases, It Helps to Blame the Market When companies raise prices, they usually tell their customers it’s because costs went up or quality improved. New research suggests that a different explanation may do more to keep customers from walking: The market made us do it. Researchers partnered with a Canadian self-storage provider in 2022 and 2023 to test how messages about price increases affected customers’ real behavior. In the study 1,626 customers received price-increase emails after they’d been with the provider for seven months. The emails used one of three justifications—cost, market conditions, or quality improvements. The study also included a no-justification price-increase message and a no-price-increase control group to capture “natural” churn. The researchers tracked the behavior of the customers over eight months. Unsurprisingly, raising the price increased attrition overall. About 33% of customers left when prices didn’t change, whereas 44% left when they did. But the reason given mattered. Market-condition justifications produced the lowest attrition—customers’ likelihood of leaving was roughly 30% lower than it was when no reason was given. Cost and quality justifications, on average, didn’t reduce churn. That’s because price-increase letters do more than trigger fairness concerns, the researchers contend. They also shape customers’ perceptions of their options. For example, when market conditions are cited, it sends the message “Good luck finding an easy alternative right now,” which compels people to stay put. In the field data, the market message worked best with customers who had fewer nearby alternatives. In follow-up online experiments, the same market justification increased people’s sense that alternatives were scarce. Quality justifications were more complicated. In online tests with North American participants, the message “We’re improving things” boosted perceived value and fairness. But reactions were split in the field. It worked with some customers and backfired with others, netting out to no benefit on average. The researchers note that their findings apply most readily to ongoing purchases, where the cost of switching would be expensive or cumbersome. For this group, a clear market-based explanation may cushion the blow of higher prices by reminding customers they may not have better options. When It’s a Bad Idea to Apologize to Customers You might think the first rule of customer service is to apologize when something goes wrong. But if customers haven’t noticed a problem, that instinct can actually hurt satisfaction, trust, and revenue. In a field experiment with a large U.S. food-delivery platform, researchers examined what happened when the firm apologized for a small failure: food arriving late by fewer than 15 minutes. For some customers, the staff followed the company’s usual practice by calling ahead, explaining the delay, and apologizing. For others, the staff did not reach out, and the order simply arrived slightly late. The researchers found that, over the next 90 days, customers who’d received an apology were less likely to place another order at all. Those who did return placed fewer repeat orders, took longer to come back, and spent less overall than comparable customers who received no apology. They also reported lower satisfaction with the experience, less trust in the provider, and weaker intentions to recommend the company. Several follow-up experiments help explain why: Apologies not only made people more likely to notice that some aspect of the service fell short (for instance, that the actual delivery time differed from the original estimate), but also led them to interpret that deviation as a “failure.” Consumers then downgraded their judgments of service quality and the firm’s competence. The researchers say that firms should distinguish between situations where customers are almost certain to notice a failure (a missed flight, a lost shipment, a never-arriving order) and situations where they might not realize that a failure occurred. In the former, timely apologies are still advisable. In the latter, it may be wiser to simply send an update or say nothing rather than flagging a “failure” the customer might never have perceived. Go deeper: “When Apologizing to Customers Hurts More Than It Helps,” by Mason R. Jenkins, Mary Steffel and Paul W. Fombelle. CEOs Are Being Appointed at Older Ages A new study found that the age of newly appointed CEOs has risen at U.S. firms, from an average of 48 in 2000 to 55 in 2023, with similar increases among CEOs at European firms. The shift is driven not by longer tenures or later retirement. Instead, more firms are selecting leaders who have held a wider variety of roles across different firms and industries, the researchers say. “Aging at the Very Top,” by Valentin Kecht, Alessandro Lizzeri, and Farzad Saidi. A More Positive Way to Approach Networking Many people avoid reaching out to new contacts, worried they’ll be a burden. But new research found that when people were trained to focus on the upsides for advice givers, they scheduled 37% more informational interviews than those who focused on benefits to themselves. “A Prosocial Perspective on Advice Seeking and Networking: How Focusing on What Advice Givers Can Gain Motivates Advice Seekers to Reach Out More,” by Anne Burmeister and Daniel Z. Levin.

Research Roundup: AI on Org Charts, Job-Hoppers’ Skillsets, Autorenewal Subscriptions, and More

Summary.

Partner Center