What AI data buyers check before they make an offer
AI data buyers check eight things before they price a company's records: industry, staff count, years of history, which systems hold the records, whether records link across systems, who owns them, where the company is based, and whether regulated data is involved. Size, history and linked records move the price most.
What could your company's records earn? Get a range in 30 seconds. Free, no obligation.
What are the eight checks?
| Check | What buyers want | Published minimum or note |
|---|---|---|
| Industry | Specialized work: factories, engineering, law, finance, health | Ceilings from $200,000 (common code) to $1,500,000 (manufacturing) |
| Staff count | Enough people to show many kinds of work | 20+ full-time staff (one buyer); 40+ "preferred" (another) |
| Years of history | Long records of real decisions | 3+ years (one buyer); 10+ years (another) |
| Systems | Several tools: email, chat, files, tickets, CRM, ERP, CAD, code | "The more tools you share, the higher your payout" (one buyer) |
| Linked records | A ticket that points to an email, a file and a decision | Buyers pay more for whole decisions than for loose messages |
| Ownership | Records the company owns outright | Client-owned files are usually out |
| Location | US, UK, Canada, EU | Most programs list these countries |
| Regulated data | Patient, financial or export-controlled records handled properly | Often excluded or specially scoped |
Sources: published buyer program pages, listings and estimators, read 20 and 21 Nov 2026.
Why do linked records matter so much?
AI companies want to teach models how work moves from a request to a result. A message alone shows one moment. A request in email, the ticket it opened, the drawing it changed and the invoice that closed it show the whole job. One buyer's estimator asks how many tools a company uses, and another names "depth of history" and "tools connected" as what sets the payout (buyer pages, read 21 Nov 2026).
Which industries do buyers value most?
Six industries carry published ceilings of $1M to $1.5M per company (buyer listings, read 20 Nov 2026): manufacturing, hardware engineering, law firms, hedge funds and quant firms, construction, and game studios. See what company data is worth by industry.
What makes buyers walk away?
- Fewer than about 20 staff, unless the work is highly specialized
- Under 3 years of records
- Records owned by clients, not the company
- Export-controlled drawings or data in scope
- A contract that forbids sharing the records
Questions
Do buyers look at our actual data before offering?
Not at first. Indicative offers are based on your answers. Some buyers check a small de-identified sample before the firm offer, always under a confidentiality agreement.
Does company revenue matter?
Published estimators ask about staff, years, tools and country, not revenue (buyer estimators, read 21 Nov 2026).
Can old records from retired systems count?
Yes, if you can export them. Older history adds depth, which buyers value.
Sources
- Published buyer program pages, listings and estimators, read 20 and 21 Nov 2026 (buyers not named by policy).