The description of an October 9 Build With AI interview with Polyshares co-founder Ryan Lock offered a striking pitch: $1 million in revenue in the company’s first month, an average deal of about $300,000, and a 6% finder’s fee. The business licenses companies’ internal records and supplies them to AI labs. These are attributed claims, not independently verified financial results. But they explain the attraction: at those figures, a successful introduction could generate $18,000. The episode is listed here.
Meanwhile, micro1 advertises $50,000-plus company-referral rewards. Mercor has its own referral program. Finding businesses willing to license their working records is becoming a business in itself.
Sales discussions, support tickets and project revisions can reveal how organizations make decisions. Yet “data broker” now covers introducers, license buyers, marketplaces and suppliers building training material. Their costs and risks differ considerably.
The investment question is whether today’s commissions buy a lasting advantage, or pay intermediaries to solve a temporary shortage of access.
An old business finds a new buyer
Information intermediaries existed long before AI. One ancestor of the modern business-information industry was the Mercantile Agency, founded by Lewis Tappan in 1841 and later part of Dun & Bradstreet. Commercial credit information helped merchants assess businesses beyond their immediate knowledge. The Library of Congress traces that history.
Consumer profiling became another branch. A 2014 Federal Trade Commission study examined nine brokers selling information for marketing, fraud prevention and other uses, while highlighting how little consumers knew about the trade.
Enterprise records licensed for AI training constitute a different transaction, although they can also contain personal information. What a buyer wants may be the sequence of decisions behind a completed job. Permission to access a company’s systems does not automatically settle permission to use everything inside them.
The economic reason for an intermediary remains familiar: information is scattered, difficult to judge and costly to prepare for another party’s purpose. Buyers face search costs; sellers may not know who wants their records or what makes them useful.
A broker can reduce those costs through relationships, qualification and predictable delivery. The challenge is preserving that contribution after the first introduction, when buyer and seller already know each other.
For paid readers: how referral fees translate into cash, what U.S. and Chinese suppliers reveal about margins, and the product and technical tests I would use to assess a durable business.