Revelio Labs

What it is#

A workforce-intelligence company. Its core asset is a panel of individual position histories built mostly from public professional profiles (LinkedIn-derived), with each position classified by occupation, a 1–7 seniority scale, reported skills, and the holder's degrees. It also classifies job postings (salary, activities, AI-related roles), employee reviews, interview reviews, and WARN layoff notices. In this vault it appears mainly as an instrument that other teams use:

• Ramp × Revelio (Kharazian, Simon & Stevens, June 2026) links Ramp card spend to Revelio headcount for 21,559 US firms. Lisa K. Simon, one of the authors, is Revelio's chief economist.
• Chandar & Klein Teeselink (September 2026) use 154M Revelio records across 41 countries. That paper's appendix gives the coverage figures that travel with every Revelio number: 67% of official employment in the US and 63% in the UK, but 24% in Germany and Brazil, 7.9% in India and 2.1% in Japan, with managers over-represented about fivefold.
• AI Writing in PhD Dissertations links 70.2% of US STEM dissertation authors to a Revelio profile to read their first jobs.

Both firm-level adoption designs in the vault code junior/entry-level as Revelio seniority 1–2. That is why their opposite composition results cannot be put down to coding.

The AI Labor Market Tracker#

AI Labor Market Tracker: September 2026 by Ben Zweig, Lisa K. Simon and Caelan Wilkie-Rogers is the second monthly edition, released 2026-10-01 with data through August 2026. The first edition (the July data) is not in the corpus. It tracks five blocks with the same indicators every month. Its framing is the useful part: exposure designs capture anticipation (employers adjusting to what AI might do), adoption designs capture realized deployment, and the two "are not contradictory." That is the same rule Seniority-Biased AI Adoption: The Junior Share at Adopting Firms derives formally: adopter-vs-non-adopter gaps and occupation-level changes can differ in sign.

Every figure is transcribed from the chart's embedded Plotly data, sampled at selected months. Nothing was read off a rendered image.

Three readings the tracker's own headlines invite and the data do not support:

1. "The postings gap narrowed from −40% to −29%" is one month. The estimate sat at −39% to −40% from January to July 2026, and the August band (−22% to −35%) barely overlaps July's (−34% to −46%). The trend that is real is the longer one: −56% in July 2025 to about −40% by early 2026. That matches the exposed-occupation postings rebound Indeed reports for 2025–26 (see Exposure Taxonomy: Observed, Theoretical, Reported, Anticipated's window sign flip).
2. "90% of activity change occurs within occupations" is not an AI signal. The within-occupation share was 84–92% in every year from 2018. It is a property of the decomposition, so it says occupation-level analyses miss most change in what workers do. It says nothing about whether AI raised that share.
3. "Adopting firms grow 27% more" is an uninstrumented gap with a visible pre-trend. Adopters were already 12.1% below their October 2022 relative level in January 2020, so they were gaining on non-adopters for nearly three years before ChatGPT. The authors say adopters "were growing faster pre-adoption". This is the design Seniority-Biased AI Adoption: The Junior Share at Adopting Firms's authors discarded on their own data for the same reason.

The adoption series has one jump worth checking. New AI-integrator adopters went from 1,014 in March 2026 to 2,644 in April, a 2.6× month, and have declined every month since. The series is a count of firms posting their first "AI integrator" job, so a classifier or posting-mix change would look the same as an adoption wave. The tracker does not discuss the jump.

Evidence note. Tier corrected vendor-claim → empirical at compile. The tracker reports event-study and difference-in-differences estimates on Revelio's own data. They are not claims about a Revelio product, and the vault grades the same kind of source (Ramp's AI Index, Indeed Hiring Lab posts) as empirical. Three limits travel with it. (1) The data are proprietary, the regressions are described in one-line captions, and the exposure score is Revelio's own. (2) The tracker is published by the firm that sells the data it reports, so it doubles as a showcase for that data. (3) The appendix shows how much the instrument matters. Revelio's exposure score correlates 0.86 with the average of itself and GPTs-are-GPTs, but only 0.19–0.50 with the other published measures (Eisfeldt 0.50, Frey & Osborne 0.41, GPTs-are-GPTs 0.39, Anthropic 0.33, Webb 0.28, AIOE 0.19). On the GPTs score the headcount gap is −10.4% rather than −6.7%, and −11.5% rather than −5.7% for workers outside 22–25.

Connections#

• Seniority-Biased AI Adoption: The Junior Share at Adopting Firms — the tracker's adoption block is a third Revelio-based adoption design, with the adoption measure that paper follows (Hosseini Maasoum & Lichtinger's postings-based AI-integrator flag) applied to US firms. It finds the senior tilt that paper finds and the pre-trend that paper warned about
• Firm AI-Spend Intensity and Headcount Growth — the other Revelio-based adoption panel, co-written by Revelio's chief economist. The tracker calls its own results "broadly similar". They are similar on total headcount and opposite on composition
• Ramp — the payment-rail partner in the June 2026 paper; Revelio supplies the workforce side
• Codified vs Tacit Knowledge Exposure — home of the tracker's exposure-design results, a second workforce instrument after ADP for the age gradient
• The AI-Exposure Pay Premium: Advertised Offers vs Realized Pay — the tracker's salary block, which points the other way from Indeed's

Sources#

• AI Labor Market Tracker: September 2026 — Zweig, Simon & Wilkie-Rogers (Revelio Labs), AI Labor Market Tracker: September 2026, released 2026-10-01 (JSON-LD says 2026-09-01; the page banner and series metadata say October 1, and the data run through August), empirical (tier corrected from vendor-claim). Cited for every block. Figures 1.1–5.3 were transcribed from embedded Plotly data at sampled months. Figure 1.2 and Figure 3.5 are mapped to their chart files by page order, because the file names are swapped. Figure 4.3 (the occupation search) is not transcribed
• A New Look at AI's Impact on Jobs: Firm-Level AI Spending and Workforce Adjustment — Kharazian, Simon & Stevens (June 2026, empirical). Cited for Lisa Simon's affiliation (chief economist, Revelio Labs) and the shared seniority coding
• How Does AI Change Labor Demand? Evidence from 41 Countries — Chandar & Klein Teeselink (2026-09-20, empirical). Cited only for Appendix A8's coverage figures; full treatment on Seniority-Biased AI Adoption: The Junior Share at Adopting Firms