Ramp and Revelio Labs pulled data from 21,000 US firms and found the opposite of the layoff...
A joint analysis from Ramp and Revelio Labs, drawn from spending and hiring data across 21,000 US firms, cuts hard against the story that companies buying the most AI are quietly gutting their org charts. The finding: heavy AI adopters grew headcount 10 percent over the past two years, and their entry level roles grew 12 percent. The firms leaning hardest into AI are also the ones hiring the most juniors.
The mechanism here is worth naming clearly, because it is the same one computing has ridden since the 1970s. When the cost of a task falls, latent demand becomes actual demand, and total work expands rather than contracts. On top of that, more AI generated output means more humans needed to review it, supervise it, and stitch it into real workflows. This is complementarity, not substitution, and it shows up in the payroll data before it shows up in the discourse.
The subtler finding sits underneath the headline number. These firms are not just hiring more juniors, they are selecting for a different candidate profile. They want people who know how to drive AI tools well, and recent graduates tend to clear that bar faster than mid career hires who have a decade of muscle memory in the old workflow. That inverts the usual assumption about seniority premiums, at least at the entry point.
For anyone building internal tools, this reframes what you are actually shipping. Your AI systems are not headcount reduction devices, they are hiring leverage devices. Teams that wire up solid agent scaffolding, evals, and observability get to hire more juniors, because each one now produces closer to senior level output. The investment in tooling compounds through people, not around them.
There is a real risk on the other side of this, and medicine is already raising the alarm. Clinicians are warning about what they call never skilling, trainees who lean on AI so heavily during their formative years that they never build independent judgment. The leverage cuts both ways. A junior who uses AI to punch above their weight while still developing their own reasoning is an asset that compounds for a decade. A junior who uses AI as a substitute for learning is a liability that gets exposed the first time the tool is wrong in a subtle way.
The firms worth watching over the next couple of years are the ones treating juniors as amplifiers rather than costs, and building the review culture that lets AI assistance accelerate skill development instead of replacing it. That is where the compounding actually happens, and it will separate the companies that got the leverage right from the ones that only got the headcount right.
Originally posted on LinkedIn.