The fear that AI replaces workers is everywhere. The actual evidence points in the opposite direction. A major new study of 21,000 U.S. businesses found that companies spending heavily on AI grew their headcount by 10% in the two years after adopting it. Companies that barely adopted AI saw no statistically significant change at all.
What the research actually found
Researchers from Ramp and Revelio Labs published a paper in 2026 called A New Look at AI's Impact on Jobs: Firm-Level AI Spending and Workforce Adjustment. They pulled real spending data across 21,000 U.S. companies and matched it to actual workforce numbers. The results were clear.
Firms in the top third of AI spending averaged about $33.67 per worker per month on AI tools. Those companies grew their total headcount by 10.2% on average over two years after adoption.
Firms that spent little on AI? No measurable change either way.
The entry-level job numbers were even more striking. High AI adoption companies grew their entry-level workforce by 12% on average. That is the job category most people expect AI to eliminate first.
The companies spending the most on AI are also the ones hiring the most. That is not a coincidence.
Why this happens
AI does not replace a department. It removes the friction that was slowing that department down. When a sales team can follow up on every lead automatically, they close more deals, which creates demand for more salespeople. When operations can process twice as many orders with the same admin overhead, the business can take on more customers, which means more staff to serve them.
AI expands what a company can handle. Growth follows.
The businesses that stay flat are the ones that do not adopt. They are stuck at the same capacity while their competitors move faster and take on more work.
What "heavy AI adoption" actually costs
The high adopters in this study were spending roughly $33 per employee per month on AI tools. For a ten-person team, that is around $330 a month, or about $4,000 a year.
That is a modest budget for tools that can automate lead follow-up, answer customer questions overnight, draft content, and handle scheduling admin. Many businesses are already spending more than that on software that does far less.
The floor for getting started is even lower. A good first AI project for a small business can often be built and running for a few hundred dollars a month, sometimes less. The bigger cost is usually the setup time, not the ongoing tool spend.
The practical takeaway for your business
This study was conducted on large U.S. firms, but the underlying logic applies at any size. AI removes the constraint on what your team can handle. More capacity means you can serve more customers. More customers means you need more people.
The question is not whether to adopt AI. The question is whether you want to be in the group that grew by 10% or the group that stayed flat.
The good news is that getting started does not require a big bet. Pick one repetitive task your team handles every week. Automate it. Measure the time you get back. Then decide where to go next.
Not sure where to start?
Take our free AI Assessment and we will tell you exactly which part of your business is most ready for AI, and what the realistic return looks like.
Take the free AI Assessment →Keep reading: The 5 Tasks Every Small Business Should Automate With AI First →