AI Adopters Hire More — But Headcount Hides the Jobs Never Posted

Firms making the largest sustained AI investments employed roughly 10.2% more workers over the two years after adoption than comparable firms that adopted later. That figure comes from a study of more than 21,000 U.S. companies by Ramp and Revelio Labs, and it anchors the newest round of debate over whether AI creates or destroys jobs. For HR leaders, the harder question is what a headcount dashboard can't show: the positions a company never posted.

The Ramp study also found entry-level headcount at heavy AI spenders grew about 12%, with gains spanning engineering, sales, administration and customer service. Companies making smaller AI investments saw no statistically significant employment effect at all. Behavioral scientist Dr. Gleb Tsipursky, writing in HR Daily Advisor, flags the caveat: the strongest adopters were already larger, more engineering-intensive, better funded and faster growing, and statistically significant sector-level gains concentrated in information businesses rather than the broader economy.

"An HR director at a hospital, manufacturer, or insurance company should therefore resist treating these findings as a staffing forecast," Tsipursky wrote. He argues the authors' claim that the data rebuts job-loss predictions goes further than the headcount numbers permit.

What does payroll growth conceal?

Tsipursky's central distinction: a company can grow its workforce while eliminating jobs that otherwise would have existed. Picture a business that expected enough demand to hire 100 employees, automates some work, and hires 40 instead. Payroll grows, nobody gets laid off — and 60 planned opportunities quietly disappear.

Displacement, he notes, can also show up as canceled vacancies, unfilled positions after attrition, reduced hours or discontinued contractor assignments. Separate Ramp research found company spending shifted away from freelance marketplaces as AI spending rose — a signal about workers who never appear on any employee roster.

What do the young-worker numbers show?

An August 2026 Stanford analysis found employment among 22- to 25-year-olds in highly AI-exposed occupations ran 19% below where it would have been had it tracked their less-exposed peers. Researchers observed weaker hiring rather than a surge in separations, and they caution the evidence doesn't establish AI as the sole cause.

A September 2026 study of Danish administrative data added a sharper finding: firms adopting AI in 2023 posted employment growth 11% lower than non-adopters by late 2025, with slower recruitment concentrated among smaller firms. The studies examine different workers, employers and outcomes, so the results don't necessarily contradict Ramp's — but together they warn HR against generalizing from high-spending adopters' success.

There's a succession-planning cost too. If employers quietly eliminate junior openings, they weaken the pipeline through which employees develop judgment and leadership skills. That's a workforce-planning problem even when today's payroll looks healthy.

What should HR measure instead?

Tsipursky's prescription is a scorecard built against the staffing plan that would have existed without AI. Specifically, he says to monitor:

• positions approved but never posted;
• vacancies canceled and replacement hires forgone;
• changes in contractor spending;
• entry-level recruiting volumes.

He recommends breaking the results out by role, department, tenure and skill level, because headcount alone captures too little.

HR should also work with finance and operations to compare labor changes against output, service quality, revenue and employee workload. A team handling more cases with fewer new hires may be more productive — but that number says nothing about whether employees face unsustainable demands or whether early-career opportunities are shrinking.

Finally, HR needs a seat where productivity gains get allocated. Tsipursky argues employers can reinvest saved time in customer service, new offerings, employee development and internal mobility, and can redesign junior roles around judgment, communication and effective AI use rather than letting entry-level hiring vanish by default.

The Ramp findings give employers a reason to explore AI's growth potential. As Tsipursky puts it, they give HR no reason to stop asking whose work is changing, whose opportunities are disappearing, and whether the organization is preparing its next generation of talent.