Research from our own data
Every figure here is measured on what happens on PeopleSearch.im: the work-email lookups it runs, the public DNS of the employers behind them, the AI clients that connect to its MCP server, and the searches people make, beside the hiring activity of a job board run by the same owner. Nothing is estimated, every source is named, and every figure is recomputed each morning.
How often can you find someone's work email?
The share of real work-email lookups that found an address, by employer size, industry, country, role and seniority.
How do B2B companies set up SPF and DMARC?
Mail provider, SPF and DMARC adoption across the employers behind our lookups, re-read from public DNS, and the email find rate by provider.
Who is hiring professionals to train AI
AI labs now hire lawyers, doctors, engineers and linguists by the hour to train and test their models. The live count of those open roles comes from the open-roles market on AITraining.jobs, a job board run by the same owner as PeopleSearch.im, which publishes it as open data under CC BY 4.0. We read it every morning.
| Kind of work | Open roles | Median pay |
|---|---|---|
| Law, Medicine & Finance | 1,911 | $80/hr |
| Multilingual & Localization | 933 | $35/hr |
| Transcription, Audio & Speech | 650 | $30/hr |
| Coding & Software Evaluation | 494 | $65/hr |
| AI Evaluation & Rating (RLHF) | 429 | $70/hr |
| Math & STEM | 292 | $40/hr |
The figures are free to quote and republish under Creative Commons Attribution 4.0 (CC BY 4.0)(opens in a new tab) with a link back. All of them in one table: figures.csv.