AI Left Your Headcount Alone. It Came for Wages and Hiring.
Sign Up For Daily Job Market News
Every week brings a new verdict on what AI is doing to work. It will erase jobs. It will create them. It will gut entry-level, lift productivity, break hiring, fix hiring. The predictions cancel each other out. So today I want to stick with two reports that skip forecasting and measure what is already happening, plus one stunt that shows how it all feels on the ground.
Start with Revelio Labs, the workforce-data firm, which on July 28 launched a monthly AI Labor Market Tracker built on workforce records, job postings, salary data, and employee sentiment.
The first finding untangles two things people keep treating as one: AI exposure and AI adoption. Exposure is how much of a job current AI could plausibly do. Adoption is whether a company actually started using it. Sort the data that way, and the contradiction disappears. Jobs most exposed to AI have grown about 4% slower than the least-exposed jobs since ChatGPT launched, and the drop is steeper for younger workers. Meanwhile, companies that actually adopted AI grew headcount 27% more than companies that didn’t.
So when you read “AI is killing jobs,” that’s the exposure number. When you read “AI companies are booming and hiring,” that’s the adoption number. Both are true at the same time. We covered this same split from a different angle in AI’s two-track job market, and Revelio just put cleaner data behind it.
One finding should stop any employer cold: the 27% adopter growth is lopsided. Senior roles at those firms are up 31%. Junior roles up just 6%. The companies leaning hardest into AI are still hiring, but the entry-level rung, the door into those firms, is barely moving. If you’re early in your career, that gap is the whole story of your job search right now. I wrote last week about how AI keeps raising the bar on entry-level hiring, and this is the mechanism underneath it.
The work is changing faster than the titles. Revelio’s index of how much the economy’s activity mix has shifted jumped to 8.4 percentage points year-over-year in June 2026, and most of that change is happening inside existing jobs, not by one occupation replacing another. Same title, different daily work. A worker keeps the role but spends less time drafting, summarizing, and entering data, and more time reviewing what the machine produced. The share of postings heavy in AI-exposable tasks has dropped about 5 points since November 2022. Revelio’s own summary line nails it: AI is automating work, but not jobs.
The last finding is the one I’d put on a slide. AI is making a hiring process that was already broken worse. The number of postings a company needs to make one hire has climbed for six years and now sits above 5. It used to be around 1. A job goes up, hundreds of applications pour in within hours, most of them AI-polished and interchangeable, and the employer can’t separate signal from noise. So they post again. The tool that promised efficient hiring made it louder.
Recruiter ghosting is the human cost of that noise, and mentions of it in interview reviews are up about 9% since October 2022. When a recruiter gets 800 applications for one opening, people stop hearing back. The volume broke the process. Revelio is honest that the breakdown started well before ChatGPT, so AI didn’t cause it. The fair read is that AI poured gas on a fire that was already burning, and a good staffing partner matters most in exactly this moment.
AI’s First Bill Is Landing in Paychecks, Not Pink Slips
The AI-jobs debate has centered on a number that keeps failing to materialize: mass unemployment. The rate stays steady, the layoff headlines never match the projections, and everyone argues about whether the models were wrong. A July white paper from Apollo Global Management offers a cleaner explanation. The effect is real. It just isn’t landing where everyone is staring.
The Apollo economists, including chief economist Torsten Slok, did something most of the earlier research didn’t. Instead of guessing what AI could do to a job, they measured what’s actually happening, matching real usage data from the Anthropic Economic Index against BLS wage and employment numbers across 321 occupations from 2015 to 2025.
The headline: so far, AI shows up in wages, not layoffs. Workers in high-exposure jobs saw real wage growth come in 6.7 percentage points slower than comparable workers after 2023, with no detectable job losses. The authors read that as firms pocketing AI’s productivity gains by holding wages down instead of cutting staff. A raise that never comes doesn’t make the news. Nobody files a WARN notice for wage compression. A worker who would have gotten 4% and gets 2% still has a job and still feels poorer every year.
Who’s absorbing it matters more than the average, and it lands on service and lower-wage workers rather than the coders and analysts everyone assumed would go first. Service occupations (childcare workers, waiters, concierges, social workers, police officers) saw real wage growth fall 24.3% versus low-exposure peers, though the authors flag that number as a small sample and tell you to read it with caution. The bottom wage quartile dropped 10.7%, the second 5.4%, the third 4.0%. Top earners showed no significant effect. Neither did blue-collar workers, whose jobs are physical. The Dallas Fed found a similar entry-level and wage pattern that we covered in AI, entry-level workers, and wages.
Right now this hits about 5.8 million workers, roughly 3.7% of the U.S. labor force, and Apollo puts a conservative floor on the cost: $28 billion in lost labor income a year. BLS projects the high-exposure workforce grows to 5.9 million by 2032. Sit with one number, though. 5.8 million affected today is far smaller than the theoretical studies predicted. The disruption people forecasted is arriving quietly, in raises that never came rather than pink slips, and if the pattern holds as adoption spreads, we’re looking at an inequality story before we’re looking at an unemployment one.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
A Tattoo for an Interview, and Why the Internet Believed It
The closer is lighter, but it lands on the same nerve. Per a Wall Street Journal report, an early-stage AI startup called LemonLime turned a San Francisco networking party into a hiring stunt. Co-founder Jordan Zietz brought a tattoo artist and offered a deal: get a tattoo at the event, get an instant interview. Seven people took him up on it. LemonLime is a five-person company in the current Y Combinator batch.
The internet turned fast. One post that went viral: “The job market cannot be this cooked that kids have to get permanent tattoos to get interviews.” Former Zynga CEO Mark Pincus, per the Journal, summed up the pushback: “It’s novel, but I don’t like it. You’re asking somebody to get the tattoo before we even know if we’re going to hire them. That’s not OK.”
To his credit, Zietz walked it back with a real apology, not a PR statement. “I messed up,” he wrote, citing “the pressure and power dynamic created by connecting tattoos to hiring.” He clarified that nobody was asked to get the company logo, the seven chose their own designs, interviews were open to everyone regardless, and he’d cover removal costs for anyone with second thoughts. It’s more accountability than most cleanups offer.
The stunt worked exactly as designed. We’re talking about a company most people hadn’t heard of a week ago, now in the Journal. In an attention economy, that’s the whole play. What touched a nerve has nothing to do with tattoos, though. It’s the power dynamic Zietz named himself. Asking someone to permanently mark their body before you’ll even talk to them, in a market where people fire off hundreds of applications and hear nothing, reads as an employer treating desperation as a filter.
The real signal is the crowd reaction, not the stunt. When “get a tattoo for an interview” is plausible enough that thousands of people believe it reflects reality, that tells you how the labor market feels to the people in it. It ties straight back to the two data stories above. Whether the pain shows up as compressed wages or absurd hiring hoops, the throughline is the same: workers feel like they’re giving up more and more just to stay in the game. The belief, real or not, is what shows up in the AI job-search trust gap we covered last week.
What This Means If You’re Hiring
Put the three together, and the picture is clear. Your org chart stays about the same size while AI compresses wages, thins the entry-level rung, and buries your recruiters under application volume it made trivial to produce. The postings-per-hire number above 5 is the one that costs employers real money and real time, because every extra posting is a role sitting open while good candidates get lost in the noise.
Solving that is the whole job. When a hiring process gets flooded with plausible-looking applications, the value is a person who can tell a strong candidate from an average one and send you only the people you’d actually hire. If your team is drowning in applicants and still not filling roles, let’s talk about how we hire for you and what our staffing services can take off your plate. And if you’re one of the workers feeling the squeeze this data describes, our open jobs are a good place to start.
