In a single week in October 2026, The New York Times, CNBC, CIO, and Moneycontrol each used Revelio Labs data to answer a different question about work. One story was about racial gaps in unemployment. One was about how companies should design teams around AI. One was about the future of call centers. One was about whether college is still worth the debt.
The topics have little in common, yet each one drew on the same dataset. That range is the point of this post: when workforce data is built carefully and updated often, it can speak to economic policy, corporate strategy, and personal financial decisions at the same time.
Here is what each story found, what the data contributed, and where our own recent research goes on.
The New York Times: 2,600 eliminated D.E.I. roles and a weaker job market than the headline rate suggests
The Black unemployment rate rose to 7% in September 2026, its fastest one-month increase since 2012 outside the pandemic, and almost double the 3.6% rate for white workers. In Black Unemployment Hints at a Job Market in Worse Shape Than It Seems, Talmon Joseph Smith argues that this gap is an early warning for the broader labor market, which looks healthier at a 4.2% overall rate than it may be.
Part of the story is about which jobs disappeared. Revelio Labs data shows that U.S. employers have cut more than 2,600 roles since 2023 with “diversity” or “D.E.I.” in the job title or description.
The Bureau of Labor Statistics (BLS) can report the unemployment rate by race. It cannot tell a reporter which specific functions employers are eliminating. That requires role-level data: job titles, descriptions, and the companies behind them, tracked over time. Pairing the official rate with that view lets the reader see both the outcome and one of the mechanisms behind it.
Our monthly jobs numbers add context to that picture. Revelio Public Labor Statistics (RPLS) shows the U.S. economy added 56.9k jobs in September 2026, with Information and Transportation losing jobs. Active job postings fell 1.8% from August to September, and 11 of 16 tracked industries declined (September RPLS report).
In the first-year RPLS retrospective, our monthly job estimates correlated with BLS at 0.77, and our postings data ran about a month ahead of the Job Openings and Labor Turnover Survey (JOLTS). That lead time matters as survey response rates fall. JOLTS response rates are now in the mid-30% range, down from about 60% in the 2010s.
CIO: the heaviest AI adopters grew entry-level headcount 12%
The common fear is that AI lets senior staff absorb junior work, so companies stop hiring junior staff. In AI won’t kill jobs. Bad team design will, Zac Engler points to research from Revelio Labs and Ramp, first reported by the Financial Times, that complicates that view.
Across nearly 22,000 U.S. companies, the heaviest AI adopters grew white-collar headcount by 10.2% over two years. Entry-level roles at those companies grew 12%.
The article is careful about what this shows. AI adopters were already larger, more engineering-focused, and more likely to be growing, so the study does not establish that AI caused the hiring. It does show that serious AI investment and a growing junior workforce coexist at scale. Engler’s conclusion for technology leaders is about team design: productivity gains will become the new baseline within a year or two, and a company that cuts its junior layer to capture them will have no one ready to step up.
Findings like this depend on company-level data: identifying which firms adopted AI, then measuring what happened to their headcount, by seniority, over the same period.
Our own research adds detail on who benefits inside those growing firms. The Revelio AI Labor Market Tracker, a free monthly publication, found that AI-adopting firms grew headcount 27% more than non-adopters since October 2022. Most of that growth was senior: senior roles rose 31%, against 6% for junior roles.
Banking shows what that looks like in one industry. Junior bank headcount is 4.7% below its October 2022 level, while senior roles kept growing at roughly their pre-ChatGPT pace. Hiring drove the decline: junior hiring fell from about 10.5% of junior headcount to about 8.5% in 2024 and 2025, while attrition returned to its normal range (Banks Are Losing Interest in Junior Talent in the AI Era). That is the risk Engler describes: fewer junior hires today means fewer people building the judgment senior roles depend on.
AI is also cited for layoffs it may not explain. Companies that blamed layoffs on AI grew AI headcount 11% over the prior two years while cutting non-AI roles more than 3%. Nearly half, though, did not fit the AI-strategy story and looked more like cost-cutting or corrections to overhiring (Is AI Just a Scapegoat for Layoffs?).
Moneycontrol: global call-center headcount is 5.3% below its peak
When a16z posted a chart showing call-center jobs shrinking about 4% a year after 15 years of roughly 4% annual growth, Elon Musk replied that they will disappear fast. The chart drew on Revelio Labs data. Moneycontrol’s Musk says the call centre is dying, and the data partly agrees tested the claim against that data and other reporting.
Revelio puts global call-center headcount 5.3% below its December 2023 peak. The decline is real, and the reporting shows it is uneven:
- The simplest tier of support, such as checking a balance or changing a flight, is shrinking fastest.
- Companies are keeping staff for complex or higher-value calls.
- The Philippines still reports call-center job growth, and employment in Egypt and Kenya kept rising after ChatGPT launched.
- Workers are mostly not being fired faster. Empty seats are going unfilled.
The article’s verdict is a slow decline rather than a sudden collapse. That is a more useful conclusion than either side of the online argument, and reaching it requires tracking the same occupation across countries and over time. For readers in India and the Philippines, where a call center is often a first office job, the difference matters.
The 5.3% figure comes from our analysis After a Decade of Growth, Global Call Center Employment Is Shrinking, which goes further than the headline number:
- Of 70 quarter-end readings since 2009, global call-center headcount fell year over year only eight times. All eight came in the last eight quarters.
- The decline began in high-income countries and moved to middle- and lower-income countries. By Q4 2025, every market was declining.
- The industry is shrinking because hiring slowed. In 12 middle-income countries, the entry rate fell from 25.1 to 17.2 per 100 workers since November 2022, below the exit rate of 18.5.
- Of call-center workers who changed jobs, 49.7% moved sideways into other customer service, sales, or office roles. Only 10.8% reached technical support, customer success, or software and data, where pay is higher.
We cannot observe which firms replace workers with chatbots. The timing, with call centers stalling after ChatGPT’s launch while other office work held up, makes AI’s role hard to dismiss.
CNBC: entry-level postings are down 35% since January 2023
CNBC’s College was ‘good debt’ for this CFP. Will it be for her daughter? by Jessica Dickler follows Stephanie Thomas, a certified financial planner in Raleigh who paid off about $20,000 in student loans and is now working to keep her daughters’ college costs down. Cost is half of the question. The other half is whether a degree still leads to a good first job.
Revelio Labs found that U.S. postings for entry-level positions fell 35% since January 2023, with AI playing a big role. That single number changes the math on a family’s biggest financial decision. The return on a degree depends on the job market a graduate enters, and postings data shows that market as it moves, well before it surfaces in graduate employment surveys.
The story also cites a separate Indeed report showing why occupation detail matters: technology and software roles are under pressure, while home health and nursing are more insulated. When a family chooses a major, it's betting on a specific occupation, and the data needs to be specific enough to inform that bet.
Our recent research shows what new graduates face when they apply. In tech, employers now ask for fewer skills but more experience. Across 75 million tech postings since 2024, the average number of required skills fell from about 30 to 21, while required experience rose from about 4.6 to 4.8 years (Tech Hiring Is Getting More Specialized). For entry-level candidates, that means competing on years of experience in the specialized skills that remain.
Hiring has also become harder to get through. The AI Labor Market Tracker found that it now takes more than 5 postings to produce a hire, up from roughly 1 or fewer historically. Those hiring trends predate ChatGPT, so we can't attribute them entirely to AI. For a family weighing tuition against outcomes, that is the market a degree has to pay off in.
The audiences are just as varied: economists and policymakers, technology executives, workers in India and the Philippines, and parents planning for tuition. Labor markets touch all of them, and good measurement serves all of them.
That measurement depends on how the data is built. Revelio Labs draws on about 2 billion career profiles and 5.6 billion job postings from 1.3 million unique sources. COSMOS, our job postings dataset, removes duplicate postings across sources. Job Architecture standardizes job titles into occupation clusters, so the same role can be compared across companies that label it differently. The data is then reweighted against government benchmarks to correct for who is and is not represented online. The data dictionary explains each step.
The data also has limits, and the best reporting acknowledges them. The CIO piece notes that the AI adoption study does not establish cause. Moneycontrol weighs Revelio’s figures alongside Goldman Sachs and Bloomberg reporting.
Workforce data is most useful when it is read this way: as precise evidence about what changed, for whom, and over what period.
Two public releases are free to follow. RPLS publishes U.S. jobs numbers each month, the day before the official BLS report. The AI Labor Market Tracker measures AI’s observed effects on hiring, employment, and the content of jobs. New analysis appears weekly on Revelio Labs News.
For company-level analysis like the work cited above, the Revelio Terminal brings workforce data across companies, jobs, skills, and geographies into one place, and the full datasets are available for delivery to research and data teams. Talk to an expert or start a free trial.