New AI labor tracker finds work changing within jobs

Revelio Labs launched its AI Labor Market Tracker on July 28, creating a recurring monthly view of how artificial intelligence is affecting the U.S. workforce. Rather than reducing the debate to jobs gained or lost, the tracker follows labor supply, employer demand, employment and wages, the activities inside jobs, and the process that matches candidates with employers.

The first edition draws on Revelio's proprietary collection of public professional profiles, job postings, employee reviews and other workforce records. Its headline results do not support a simple automation story. Since October 2022, demand for the most AI-exposed occupations has fallen 42 percent relative to the least exposed group, with the decline concentrated in junior roles. Yet headcount at companies identified as AI adopters has grown 27 percent relative to non-adopters over the same period.

Those findings are not necessarily contradictory. The tracker distinguishes exposure—whether AI could technically perform activities in a role—from adoption, where job descriptions and profiles show that an organisation is actually using AI tools or workflows. Companies may cut hiring in work they expect AI to affect while expanding after successful deployment creates new capacity or demand. The data does not prove that AI caused every difference.

Workers are responding as well. Computer Science and IT enrolment in the measured sample has fallen 28 percent from its 2022 peak, while 31 percent of newly reported certifications in June 2026 were AI-related. Almost half of those AI credentials concerned generative AI and large language models. That combination suggests rapid reskilling alongside uncertainty about traditional education routes.

The most revealing signal may be inside occupations. Revelio says the mix of activities performed across the economy shifted sharply during the past three months, with most change happening within existing roles rather than through wholesale replacement of occupations. At AI-adopting firms, business-outlook sentiment improved relative to non-adopters, but perceived job security and confidence in senior leadership weakened, and layoff anxiety increased.

For professionals, the practical message is to watch how tasks are changing, not only whether a job title survives. For employers, the tracker offers a reason to connect AI investment with hiring, training and employee sentiment instead of measuring software usage alone. Its limits matter: this is a U.S.-focused, observational product built largely from public and proprietary data, not an official labour statistic or a causal verdict. Monthly updates should make it more useful by showing which early signals persist.