Companies that spend the most on generative AI are hiring more people, not fewer. That is the finding from a joint study by Ramp and Revelio Labs, which tracked firm-level AI spending against employment outcomes across more than 21,000 US companies. Two years after adopting AI at high intensity, these firms grew total headcount by 10.2 percent and entry-level hiring by 12 percent.
The result cuts against the dominant narrative of the last two years: that generative AI tools would hollow out junior roles first, since entry-level work is disproportionately made up of the tasks large language models are best at automating. Ramp’s data says the opposite happened at the companies spending the most on AI. Entry-level hiring did not just hold steady. It outpaced overall headcount growth.
The methodology matters here. Ramp has direct visibility into corporate card and expense data, which gives it a real-time proxy for AI spending that survey-based studies cannot match. Revelio Labs supplies the employment side, drawing on resume and job-posting data across a large employer sample. Pairing transaction-level spend with workforce records is a more grounded approach than asking executives to self-report their AI adoption plans.
None of this proves AI adoption causes hiring. Companies with the cash and confidence to spend heavily on AI tools are also the companies growing fastest for other reasons: strong revenue, market share gains, expansion into new lines of business. The correlation Ramp found is consistent with AI spending as a signal of growth, not necessarily the cause of it. The study does not appear to isolate that distinction, based on the available summary.
Still, the direction of the finding is hard to dismiss. If AI were primarily a labor substitute at these companies, headcount growth would show up concentrated in senior roles while entry-level hiring lagged or shrank. Instead entry-level hiring grew faster than the company-wide average. That is closer to what a growth complement looks like: AI increasing the output a company can support, which increases the number of people needed to sell it, service it, and build on top of it.
The caveat is time horizon. Two years is long enough to capture initial hiring waves following AI adoption, but it is not long enough to capture what happens once a company has fully redesigned workflows around AI tooling rather than layering AI on top of existing headcount plans. The layoffs some feared may simply arrive on a longer clock than the one this study covers.
For now, operators weighing AI investment against headcount plans should treat this as evidence, not proof, that the two move together rather than in opposition. Any team building a business case for AI spend that assumes it will let them shrink headcount should test that assumption against a specific role and task, not the industry mood.
Reported by Ramp on July 2, 2026.
