Is my company too small to license its data to AI buyers?
· Company Data Advisors
Most AI data buyers with a size floor set it at 20 to 50 employees, but law, accounting and medical firms qualify smaller, and nine programs state no minimum.
Maybe not. Of the 17 AI data-buyer programs in our buyer index, the ones that publish a general size floor set it between 20 and 50 employees (two of them flexible), and a company under 20 is below every one of those general floors. But the floors are lower for some kinds of firm: law, accounting, investment and medical practices can qualify with 5 to 10 people at some programs, and nine programs state no minimum headcount at all. Our read: size is a stand-in for what buyers actually want, which is several years of work by a team, recorded across several linked systems.
We read every program's own page on November 13, 2026. Floors change, so check the page before you apply.
What are the minimum company sizes for AI data buyers?
AI data buyers' size floors range from "any size" to 50 full-time employees at peak. The table sets out what each program in our buyer index says on its own page.
| Program | Size it states | History it states |
|---|---|---|
| SourceX | 50+ full-time employees at peak, contractors excluded | Several years of operating history |
| Collinear | About 50+ people is the best fit; 15+ at peak raises the price; smaller teams with deep, specialized records can work | Several years of history, written by people |
| Scale | 40+ employees is its "preferred starting point", assessed case by case | Years of records in the tools your teams use |
| micro1 | 30+ full-time employees | Several years of records across several systems is a stronger fit |
| Mode | 6+ for legal; 10+ for accounting, finance and medical; 30+ for all other businesses | Several years of records the company owns |
| Handshake | 20+ full-time employees | 3+ years in operation |
| Nyne | 20+ full-time U.S. W-2 employees | Three years in business |
| Replay | No minimum stated; its lowest published pay tier is 20 to 50 employees | How long you've operated affects price |
| Mercor | Set per listing, from "Any size" (imaging centers) and 5 (six listings) up to 50 | Not stated on its listings page |
| SuperAnnotate | "Any size" | Records built up over time |
| Polyshares | No minimum; the record matters more than headcount | Real operating history |
| Labelbox Prism | Not stated | Not stated |
| Turing (Project Lazarus) | Not stated | Not stated |
| Centific | Not stated | Not stated |
| Appen | Not stated | Longer records are worth more |
| Protege | Not stated | Not stated |
| The General Data Company | Not stated | Not stated |
Sources: each program's own page, read on November 13, 2026. Each buyer's page in our index links to its program.
Programs at the top of the table are the strictest. SourceX says it is "best suited to companies with 50+ full-time employees at peak (contractors excluded)," and its November 11 announcement repeats that "the strongest candidates had 50 or more full-time employees at peak." Scale prefers 40 or more and says it assesses "every partnership on a case-by-case basis." micro1 asks for "30+ full-time employees." Handshake and Nyne both ask for 20 full-time employees and three years in operation.
Why do AI data buyers set a minimum company size?
Our read: AI data buyers set size floors because a small team rarely produces enough varied, connected work to build what they sell. Several buyers, including Collinear and Turing, turn company records into practice environments where an AI agent attempts a task and is scored, which our post on flight simulators for office work explains. That needs many examples of people handing work to each other across tools.
Reporting and buyers' own pages point the same way. The New York Post reported on November 10 that about 200 businesses a day sign up to explore licensing with micro1, and that "the smallest ones rarely make it through the process because they simply don't produce data with enough volume or complexity to make it worth buying." micro1's CEO told the paper that most businesses it works with have between 30 and 150 employees. Collinear's best fit is "teams of about 50 people or more, with the same work visible across several tools," and it lists "code from very small teams with little history" among what it does not need. Replay prices partly by headcount but adds: "Headcount is a proxy. How long you've been operating, what industry you're in, which apps you run on and how much data sits in each all factor into the final price."
Our view: treat the floor as a test of whether your records show a team at work over time. A 15-person firm with ten years of tickets, email and project files that link to each other is closer to what buyers want than a 40-person company that started using its systems last year.
Which AI data buyers accept small firms?
Among programs that set floors by industry, professional firms get the lowest. Mode sets its minimum at "6+ for legal; 10+ for accounting, finance, and medical; 30+ for all other businesses or nonprofits." Mercor publishes 26 referral listings, each with its own size range:
| Mercor listing | Size it states |
|---|---|
| Imaging & Diagnostics Center | Any size |
| Law Firm | 5 to 300 attorneys and staff |
| Accounting & Tax Firm | 5 to 200 employees |
| Investment Firm | 5 to 200 employees |
| Hedge Fund or Quant Firm | 5 to 500 employees |
| AI Product Company | 5 to 200 employees |
| Video Production Studio | 5 to 500 employees |
| Medical Practice or Health System | 10+ clinicians or 2,500+ patients |
| Consulting & Advisory Firm, Design & Web Agency, Game Studio, and five others | 10 minimum (upper limits vary) |
| Software Company | 15 to 500 employees |
| Google Workspace Power User, Manufacturing, Hardware Engineering, Cybersecurity | 20 minimum (upper limits vary) |
| Construction, SaaS & Subscription (with a data team), Call Center or BPO (agents) | 50 minimum (upper limits vary) |
Source: Mercor's listings page, read on November 13, 2026. Two listings, payments and chemistry R&D records, state no size.
A small law, accounting or medical practice therefore has more options than a small company in most other fields. Each of those fields also has its own rules on client and patient information, covered in our guides for law firms, accounting firms and medical practices.
Nine programs publish no headcount minimum. Replay states none, though its lowest published pay tier starts at 20 employees. SuperAnnotate says "Any size, any industry, anywhere," and adds the condition: "You need records built up over time and the rights to license them." Polyshares says "the length and texture of the record matter more than headcount or revenue." Labelbox Prism, Turing, Centific, Appen, Protege and The General Data Company state no size rule on their pages. No stated floor doesn't mean no review: each still decides whether your records fit what it needs.
How do buyers count employees?
Most AI data buyers don't say how to count, and the ones that do count differently. Nyne counts "full-time U.S. W-2 employees." SourceX counts "full-time employees at peak," and its fit check notes "Not counting contractors." Collinear speaks of "people at peak." micro1 and Handshake say full-time employees without saying when. Mercor's law firm listing counts "attorneys and staff," and its medical listing counts clinicians or patients.
Peak headcount matters for two kinds of company: those that have shrunk, and those that have closed. SourceX says companies "that are still operating, acquired, or wound down can all qualify," measured at peak. Turing takes companies that are winding down or wound down. Centific and Appen both answer companies that are shutting down, and Centific's form lists "Already closed" as an option. A company that had 60 people in 2021 and has 12 today may still clear a peak-based floor, because the records from the larger years are what a buyer would license. Our guide to licensing when you're closing the business covers who can sign once a company is winding down.
Our view: give each buyer the number its page asks for, and say how you counted. If you rely on contractors, part-timers or staff outside the US, say so on the form rather than rounding up. A buyer checks the records against the description, and a headcount that doesn't match the records costs you credibility before the offer stage.
How many companies clear these floors?
The average U.S. employer firm has 11 to 15 employees, below a 20-employee floor. The SBA's Office of Advocacy counts 36.2 million small businesses, of which 6.4 million have paid employees, and reports that in 2022 family-owned employer firms averaged 15 employees and other employer firms 11. Our view: with averages that low, the typical employer firm sits below every general floor in the table.
That is why the exceptions matter. If you are a small professional firm, start with the programs that set industry floors. If you are a small company in another field, start with the programs that publish no floor, and expect them to judge the depth of your records rather than your headcount.
What can a small company do?
A small company's best route depends on why it is small. Here are the cases we see most:
- You are a law, accounting, investment or medical practice. Start with Mode's industry floors and the Mercor listing for your field.
- You were bigger at your peak. Ask the programs that count at peak, and describe the years when you were largest.
- You have a small team but deep, specialized records. Collinear says "smaller teams with deep, specialized records can work too," and Scale assesses case by case. Describe the years of history and the systems that link to each other, not just the headcount.
- No program fits today. Keep your records. Several buyers stress that history disappears when subscriptions lapse, and a company that grows past a floor in two years will want those years of records intact.
Being under a floor is not the same as having nothing of value, but it narrows the field, and some small companies won't fit any program today. Our post on whether your company's data is worth anything covers what raises or lowers interest beyond size.
Where we come in
In our experience, a small company fits fewer programs than a larger one, and the rules that decide fit (industry floors, peak headcount, contractors, history) are spread across 17 different pages that change often. We track every program's published criteria and terms, have direct contacts at the major data buyers and at AI labs, and can tell you which programs your company fits before you fill in any forms. If none fits, we say so. Nothing about your company is shared until you approve a specific, named buyer. Participating buyers may pay us referral fees, and we tell you which before any introduction. Owners pay us nothing, and you can apply to any buyer directly.
What to do next
Before you apply anywhere, write down four numbers and two facts:
- Full-time employees today, and the number that are U.S. W-2 employees.
- Full-time employees at your peak, and the year.
- Contractors and part-timers, counted separately.
- Years of records in your main systems (email, chat, tickets, CRM, accounting, project files).
- Your industry, in the words buyers use (law firm, accounting firm, medical practice, software company).
- Whether the company is operating, winding down or closed.
Then compare those against the table above, or check where you fit. For the steps after that, from first form to payment, see how to sell or license your company's data to AI companies.
Programs in this post
More from the blog
- What company data AI buyers want: November 2026
Across the 17 AI data-buyer programs we track, accounting, healthcare and software lead the industries, and CRM, ticket, email and finance records lead.
- Can a software company license its code to AI data buyers?
12 of the 17 AI data-buyer programs we track name software or code. They want private code with its history. The code is often yours to license; customer data often isn't.
- How to check that an AI data buyer is legitimate
Six checks an owner can run in an afternoon before sharing records. Who the buyer is, what you pay, what it asks first, personal data, payment and license terms.
See which buyers fit your company
Eight quick questions against every program in our buyer index. Participating buyers may pay us referral fees.