Saving time on menial tasks so employees can instead focus on more meaningful work is often trumpeted by artificial intelligence boosters
But, a
Study co-authors Xingqi Maggie Ye, a doctoral student, and associate professor Aruna Ranganathan found that AI tools helped employees work faster, take on more tasks and extend the workday without being asked to do so.
"That may sound like a win, but it's not quite so simple," they wrote. "These changes can be unsustainable, leading to workload creep, cognitive fatigue, burnout and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover and other problems."
William Trout, director of securities and investments at technology data firm Datos Insights, said if anything, these findings undersell the problem. He said what he calls the "parallel-threading effect" is something he observes constantly in wealth management.
"People running multiple AI agents at once, picking up tasks they'd shelved for months because now there's something to offload them to," he said. "It feels like momentum. It feels like finally having a capable collaborator."
What's actually happening, Trout said, is that advisors' attentions are fractured across a dozen open loops simultaneously.
"You're not doing one thing well," he said. "You're supervising several things at once, and that supervision is its own cognitive tax."
Experts say it's up to firm leaders to carefully scrutinize how employees are using AI tools and set boundaries to prevent burnout.
READ MORE:
Is any time actually being saved?
In his experience, Trout said
"You can do more — faster, at higher quality, across a wider range of tasks," he said. "But 'more' tends to expand until it fills whatever time you freed up, and then some."
John O'Connell, CEO of The Oasis Group, said AI tools transition time from administrative tasks to higher-value tasks, such as client engagement.
"At best, they can reduce the overtime spent by advisory firm team members 'catching up' on paperwork from their day, leading to overtime," he said.
Time saved on a task is real, but capacity without intentional direction gets consumed immediately by the next thing, said Jerry Robert, head of data and AI at F2 Strategy in Chicago.
The question firms need to be asking isn't, "How much time did AI save us?" said Robert, it's, "What are we deliberately doing with that capacity?"
"Without an answer to that second question, you haven't saved time," he said.
"I'm doing more, faster, at a higher standard than before," he said. "The productivity gain is real, but it quietly resets your baseline expectations rather than reducing your workload."
When you remove friction from a task, people don't protect that reclaimed time — they fill it, said Robert.
"That's entirely predictable human behavior, and it's exactly what change management is supposed to get ahead of," he said.
The people who feel liberated by AI are usually the ones setting the agenda, said Trout.
"The people doing the execution are the ones quietly working longer hours and wondering why they're exhausted," he said.
As an advisor who uses AI tools in his practice, Benjamin Simerly, founder of
"To those employees, AI will be another tool in the box," he said.
READ MORE:
How should firms change expectations?
The first thing Trout said he tells firms is: don't mistake acceleration for sustainability.
"The early months of AI adoption almost always look great — output is up, energy is high, people are engaged," he said. "But you're often just borrowing against future capacity."
What firms actually need to do, Trout said, is treat AI adoption like any other operational change: map the workflows, define what good output looks like and set explicit norms around when to stop.
"Otherwise the standard just keeps drifting upward, and the people doing the work keep running to catch it," he said.
The core discipline is treating capacity savings as a resource to be allocated, not a byproduct to be absorbed, said Robert.
"The firms that get this right govern the capacity, not just the tool," he said.
Robert said his specific recommendations to firms are to:
- Name the redirect: When AI saves time on a task, explicitly decide where that capacity goes before it disappears into the workload.
- Set scope boundaries: Define what AI-assisted work looks like when it's done, so there's no drift into endless expansion.
- Build review checkpoints: Intentional pauses to assess quality, not just volume, before moving to the next thing.
- Make managers accountable for workload norms, not just output targets: Burnout shows up in their data before it shows up in turnover.









