AI Is Making Workers Into Managers, But Paychecks Haven’t Caught Up
A growing share of employees are quietly taking on management duties, not over people, but over AI agents, and finding that neither their job titles nor their salaries reflect it.
Across industries, workers who once spent their days coding, writing, or designing are now spending a large chunk of that time directing artificial intelligence instead: setting up tasks, checking outputs, fixing errors, and deciding when a piece of work is good enough to ship. The work looks a lot like management. The compensation does not.
The numbers behind the shift
A worldwide study of nearly 12,000 employees by the Boston Consulting Group found that almost half, 47%, now spend more time overseeing AI systems than actually doing the underlying work themselves. Two-thirds said AI had absorbed the simpler parts of their jobs, leaving them with the harder, more judgment-heavy tasks.
Yet the support hasn’t kept pace with the change. Roughly seven in ten workers in the same study said AI had meaningfully reshaped the skills their job demands, but only about a third said they’d actually received training to match.
Engineers, writers, marketers all affected
The trend shows up across very different roles. A software engineer at a major appliance company described managing several AI coding agents at once as its own kind of loop: setting parameters, watching progress, and deciding whether to let a task run again or call it finished. She said the work is less about solving the technical problem directly and more about coordinating the process around it.
A freelance technical writer estimated she now spends up to 30% of her time supervising AI tools rather than writing herself, calling agent management a skill she had to develop out of necessity. A marketing consultant reported a similar share of her time going toward overseeing AI “agents” that she treats almost like team members, complete with names and defined responsibilities.
Even outside the tech and marketing world, the shift is landing on people who never signed up for it. One university instructor said she’d been put in charge of managing AI use in her classroom despite having no training in the technology and real reservations about it, caught between telling students to embrace AI and warning them not to lean on it for their coursework.
Titles and pay lag behind
Separate research from workforce analytics firm Visier found that more than half of full-time U.S. employees say AI has substantially changed their role in the past two years, but only a small fraction of that group, under one in five, has actually been promoted as a result. Researchers who study the data point out that people who formally manage other humans still earn meaningfully more than individual contributors doing comparable work, and that gap hasn’t closed just because someone is now managing a fleet of AI agents instead of a team of people.
One executive at a compensation software firm summed up the disconnect: if a worker is assigning tasks to multiple agents, judging their output, and taking responsibility for what gets delivered, it’s fair to ask whether “individual contributor” is still the right label. For now, most companies haven’t figured out how to answer that question. A few are experimenting with bonuses tied to effective AI oversight, but broad increases to base pay don’t appear to be on the table yet.
The bottom line
Companies are layering AI responsibilities onto existing jobs and existing pay structures, largely without redesigning either. Employees are absorbing new managerial-style work, and in many cases figuring out how to do it well on their own, while their titles, training, and compensation remain stuck in place.
Source: Business Insider
