New research suggests AI adoption may be driving hiring growth, not job losses
The narrative around AI and jobs has been shaped largely by layoff announcements and displacement warnings. Oracle cuts 21,000 positions. Snap, Cisco, and Block cited AI when announcing layoffs. A new graduate is told to learn AI and then, in the next breath, warned that AI will cost them their job.
But new research suggests the picture may be more complicated.
Ara Kharazian, lead economist at Ramp, a San Francisco-based AI finance platform for businesses, is co-author of a 2026 study that takes a novel approach to measuring AI's effect on employment. Produced with workforce data firm Revelio Labs, the research tracked nearly 22,000 U.S. companies, linking Ramp's corporate spending records to Revelio's workforce data compiled from public professional profiles.
"We'd have a data set that shows, hey, these firms spent on AI, these firms didn't, and then have their headcount change over time," Kharazian said. "That's where we see the really interesting results."
AI spend and hiring growth
Among companies that adopted AI and used it most intensively, white-collar headcount grew 10.2% over the two years following adoption. Entry-level hiring jumped 12%. Among light adopters, there was no statistically significant change.
That's because, for many companies, AI adoption is still in its early stages.
"For many firms adopting AI so far, their adoption of AI is pretty minimal," Kharazian said. "They may have a chatbot subscription here and there, but it's not particularly integrated into their workflows and it's not particularly productivity enhancing."
On the opposite end of the spectrum, the study defined high-intensity users as companies in the top third of AI spending per employee per month. In practice, that threshold is lower than it might sound. High-intensity firms spent an average of about $30 per employee per month in their first three months of adoption.
Along with spend, the gap between casual adopters and high-intensity users comes down to depth of integration. High-intensity companies use multiple AI models and invest in more advanced tools like coding agents and API services, rather than a basic chatbot subscription.
The six-to-twelve month gap
Those gains don't arrive immediately. Companies that adopt AI typically don't see headcount gains for at least six months to a year, according to the study. He points to two likely reasons. Workforce composition changes take time to work through an organization, and companies need time to figure out how to integrate AI effectively before they know where to invest.
This matters for HR leaders evaluating their own adoption. Organizations that have already struggled to connect AI spending with measurable productivity gains may simply be earlier in that curve than they realize.
Entry-level workers may be a bright spot
Perhaps the most counterintuitive finding is the surge in entry-level hiring among high-intensity AI adopters. That 12% increase runs against the prevailing anxiety about what AI means for new graduates.
The data hints that companies may be hiring differently, not less. They want workers who already know how to use AI well, and recent graduates are often better positioned than their more senior counterparts.
"We believe that they're hiring, seeking employees who know how to use AI and use it well," he said. "And what better place to look than recent grads and college students who are already quite AI native?"
This connects to a broader tension in the talent market. As HRD has reported, AI-driven hiring is reshaping how candidates are screened and evaluated, even as employers struggle to define what AI fluency looks like in practice.
Gains are real, but uneven
The study comes with important caveats. Almost all headcount gains were found among technology-sector firms, and the research covered only white-collar workers — a group already facing declining job postings and stalling wages, according to separate Revelio Labs research.
The study also accounts for a key methodological question about whether fast-growing companies were simply already on a hiring trajectory before AI entered the picture. The study's preferred analysis compares early AI adopters against calendar-matched firms that hadn't yet adopted, rather than firms that never adopted at all.
"Even when we compare firms that are growing at similar rates, their growth accelerates following AI adoption relative to firms that have not adopted yet," he said.
For organizations still figuring out their approach, Kharazian's message is that a ChatGPT subscription probably isn't enough.
"If you tried it and you found that you're not getting the productivity gains that you expected, you probably need to keep trying it," he said. "These gains are concentrated amongst firms with sustained adoption and who are using the most advanced tools available."
He also flagged a competitive blind spot. Companies using AI effectively have little incentive to document how they're doing it.
"Unfortunately, we're in this market in which your competitors, who may be using AI very productively, have no incentive to publish their playbook," Kharazian said. "So you're not going to get this kind of advice from other companies like yours."
The debate over AI's net effect on employment is far from settled. Research covered by HRD found that many AI leaders warn of displacement while simultaneously accelerating adoption, a contradiction that leaves HR leaders navigating the gap between public messaging and operational reality. What's clear is that AI's effect on employment is neither uniform nor inevitable. Where companies land may depend less on whether they adopt the technology, and more on how seriously they do.
