Firms That Adopted AI Heavily Grew Headcount 10%, New Ramp Study Finds

    A Ramp and Revelio Labs paper linking AI spending to workforce data across 21,000 US businesses found heavy adopters added jobs, while low adopters saw no change

    Firms That Adopted AI Heavily Grew Headcount 10%, New Ramp Study Finds A Ramp and Revelio Labs paper linking AI spending to workforce data across 21,000 US businesses found heavy adopters added jobs, while low adopters saw no change Aaron Rafferty June 30, 2026 Key Takeaways A Ramp and Revelio Labs study of 21,000 US firms found that heavy AI adopters grew headcount about 10% over the two years after adoption. Entry-level roles grew faster at those firms, up about 12%, while companies that spent little on AI saw no statistically significant change. The finding is a correlation, not proof that AI creates jobs, but it cuts against the idea that buying AI means cutting staff. A new study pushes back on the idea that AI is destroying jobs. Ramp economist Ara Kharazian shared a paper he wrote with Revelio Labs that links firm-level AI spending to workforce data across 21,000 US businesses, and the headline finding is that companies adopting AI heavily grew their headcount. What the data shows Firms in the top third of AI spending, which works out to about $33.67 per worker per month, grew headcount roughly 10% over the two years after they adopted AI. Entry-level roles grew even faster at those firms, up about 12%. Companies that spent little on AI saw no statistically significant change either way. The Financial Times, which covered the paper, summarized it the same way: more AI spending lined up with more workers, not fewer. What it does not prove The finding is correlation, not proof that AI creates jobs on its own. Firms that are already growing may be the ones with the budget and ambition to adopt AI, so some of the headcount gain reflects healthy companies doing what healthy companies do. The study also measures jobs inside adopting firms, not across the whole economy,

    Loading full article…