AI-adopting firms are hiring 27% faster than everyone else. Almost all the growth is going to senior roles.

Revelio Labs launched what it calls the first real-time, evidence-backed measure of AI's impact on the workforce on July 28 -- a monthly tracker built on the labor-analytics firm's proprietary data covering more than 30 million companies and 5 billion job postings. Rather than a one-off study, it is designed to be republished every month, with the next edition due mid-August. Its inaugural read, covering the period since OpenAI's [ChatGPT](#/company/openai) launched in November 2022, is the most granular public look yet at who is actually gaining and losing as employers adopt AI.

That distinguishes it from the labor-market debate it enters. Most public evidence on AI's effect on jobs so far has come from either single-company anecdotes -- a layoff announcement citing AI, a hiring freeze attributed to it -- or backward-looking academic studies published a year or more after the fact. Revelio Labs is instead standing up a live instrument: the same methodology, rerun against fresh job-posting and headcount data every month, against a fixed reference point (AI's mainstream arrival via ChatGPT) rather than a moving one. That design choice is what makes a monthly delta -- like the one due in mid-August -- meaningful, rather than just another one-off snapshot to compare against different studies with different methods.

Who's actually growing, and who's captured it

The tracker's headline finding is a divergence between companies, not just occupations: firms that have adopted AI grew total headcount 27% more than firms that haven't, since late 2022. Employment in AI-related roles specifically grew 18%, against 3% for non-AI roles at the same firms. Revelio Labs' chief economist, Lisa Simon, summed up the underlying pattern: "The content of work is changing quickly, but most of that change is happening within existing occupations" rather than by creating or eliminating entire job categories wholesale.

The tracker doesn't state why the growth has concentrated so heavily in senior roles, but the pattern is consistent with something already visible in broader labor data this year: payrolls in tech and financial-activities roles -- the sectors adopting AI fastest -- have been shrinking by an average of 28,000 jobs a month in 2026, even as overall U.S. hiring stayed positive. Read together, the two data sets point the same direction: employers appear to be consolidating headcount growth in experienced roles while pulling back on entry-level hiring, rather than cutting broadly across seniority levels.

That concentration matters beyond this year's hiring numbers, because seniority in most white-collar fields is a ladder, not a set of parallel tracks: today's junior hires are the pool a firm would normally promote into tomorrow's senior roles. If AI-adopting firms keep growing senior headcount by expanding externally -- hiring already-experienced people, rather than developing them internally -- while junior hiring stalls at 6% growth, the tracker is describing the first data-backed sign of a narrowing entry point into the exact firms otherwise expanding fastest. Revelio Labs' tracker doesn't yet measure promotion rates directly, which is the harder number that would confirm or complicate that read.

The wage data points the same direction as the headcount split. Before ChatGPT's launch, jobs in the most AI-exposed occupations paid roughly a 2% premium over comparable work -- employers rewarding skills that were, at the time, scarce and valuable. That premium has since eroded to roughly zero. Combined with the -13% employment gap for workers aged 22 to 25 in those same occupations, the tracker's picture of AI exposure is not just about whether a role exists, but about whether it still pays what it used to for the people newest to it.

The exposure numbers, and where Revelio's own sources disagree

Beyond headcount, the tracker measures hiring demand directly through job postings -- and here Revelio Labs' own materials don't quite agree with each other on the size of the drop.

That erosion sits on top of a smaller but real aggregate gap: employment in the most AI-exposed occupations is down about 4% relative to the least-exposed occupations over the same period, before splitting by age or seniority at all.

On the labor-supply side, the tracker also finds computer-science enrollment down 28% since 2022, even as workers try to reskill toward where demand is growing: 31% of all newly issued professional certifications in June 2026 were AI-related, and nearly half of those, 47%, were specifically generative-AI or [LLM](#/dictionary)-focused -- the fastest-growing credential category the tracker measures. Read against the enrollment decline, the certification numbers suggest two different populations responding to the same shift: prospective students steering away from computer science as a degree path, and people already in the workforce reskilling toward AI specifically, often faster than universities are adjusting their own program mix.

None of this amounts to proof that AI is destroying entry-level work outright -- the tracker itself frames most of the change as happening within occupations, not through wholesale elimination of job categories. What it does establish, for the first time with a live, repeatable instrument rather than a one-off survey, is that the benefits and costs of AI adoption are not landing evenly even within the same companies. The firms growing fastest because of AI are, on this data, growing mostly at the top of their own org charts.