QualifyDataQualifyDataYes — AI companies pay nonprofits to license their records for AI training. What they are buying is the operating history of a mission organization: years of program, service-delivery and administrative records. Identifying information is removed before any AI developer sees a line of it, and your organization decides what, if anything, goes.
Every case opened, every service delivered, every report filed built a history inside your systems — the history AI developers need to learn how organizations like yours actually operate. They license a cleaned, de-identified copy of those records. You keep your originals, you keep using them, and your programs keep running as they do today.
QualifyData is a broker. There is no need to submit to one buyer direct and take the one number they offer. We put your archive in front of 15 AI companies, let them bid, and submit the best offer to you, in writing. You go through the process one time.
Five minutes. See if your archive qualifies.No cost. No commitment. No logins, no records handed over — just a few questions about your archive's size.
The records that show how the work actually got done — built up over years, across the ordinary systems a nonprofit runs on:
More sources and more years make a stronger archive. You pick which systems and which years go — nothing leaves until you sign a license that lists exactly what goes.
Years of records across several systems — program management, communication, documents, accounting. A steady desk staff working in them daily. Records you control and can export. A closed or merged organization measured at its peak counts too.
Records that stayed on paper, never digitized. One system with a short history. Records you no longer control or cannot get to. A history under three years. The check sorts this out in five minutes, at no cost.
Plainly: beneficiary and client records are the sensitive core of the archive, and they are treated that way. Before anything moves, a cleaning process removes identifying information — names, emails, phone numbers and addresses are replaced (a person becomes "Person 01"), dates are generalized, and card and bank numbers and anything about minors are left out entirely. Only the buyer's cleaning process sees the raw copy; developers get the cleaned version, and the raw copy is deleted on the schedule in your license. You can ask to see a cleaned sample first — before anything is final.
And the part worth saying out loud: you do not have to include beneficiary records at all. Many organizations will rightly keep client records out entirely and license only the operations side — program administration, correspondence, grant files, accounting and HR. That is a legitimate choice, and the archive can still qualify on those records alone. Which systems go is your organization's call, made with your own counsel, who reviews every document before you sign anything. This page explains the process; it is not legal advice.
Five minutes. See if your archive qualifies.For most nonprofits, yes. Licensing an archive built over years is a board-level decision. The board sees the offer in writing before anything is signed. Your counsel can review the services agreement and the license. And whether licensing de-identified records fits your mission is the board's call, not ours.
The decision itself is narrow: a cleaned copy of records the organization already holds, identifying information removed, for a set term, originals in place. If the board's answer is no, that is the answer — nothing has been handed over. If yes, the license lists exactly what is included.
The general shape: an organization in the US, Canada, the UK or Australia, operating for 3 or more years, with 20 or more people working at desks. "Desk staff" means people whose work runs through the systems all day — program administration, case management staff, development, finance and the front office. Records should run across several systems — five or more is a strong archive.
A small organization with a handful of staff will usually be under the line. A mid-size or large organization with an administrative staff is often over it. The AI companies make the final call on fit, starting with the five-minute check.
It can still qualify — a closed organization is measured as it was at its peak, in its final stretch of real operation: the desk staff it had, the years of records it built, the systems it ran on. If it merged, the records may now sit with the successor organization. If it closed, they may sit in retained systems, in storage, or with a records custodian. Either way, they are an asset somebody may be paying to keep alive — and a license can collect that value before accounts lapse or retention periods run out.
Who holds the right to license those records depends on how the closure or merger was structured — storage is not the same as ownership. Your attorney confirms who can sign before anything moves. The sensitivity rules do not change — identifying information comes out first, and beneficiary records can stay out entirely. Our closed-companies page covers the rest of the picture.
Five minutes. See if your archive qualifies.Mostly, the small and the thin. An organization is usually not a fit if it has fewer than about 20 people working at desks, fewer than 3 years of records, an archive that lives mostly on paper, or records held in a single system with little history. Records you cannot access or no longer have the right to license cannot go either. The five-minute check will tell you quickly, at no cost.
You never hand over a login, and you see every number in writing before you commit to anything.
We walk through your systems and answer your questions — including the sensitivity and governance ones. No data, no logins, no commitment.
A client services agreement lets us represent you. It doesn't commit you to a deal, and your board and attorney are welcome on a call first.
60 to 90 minutes, system by system: how far back, how much. Counts and sizes only — no records, and nobody gets a login.
That list — no actual records in it — goes out through our broker network to 15 AI companies. They bid, and we submit the best one to you, in writing.
Your board and attorney review the terms, and you sign directly with the AI company behind that offer. The license lists exactly what is included.
Your team exports copies, with help at every step — typically 5 to 10 business days. Then a buyer review of about 30 days.
The AI company pays you the license fee directly.
From signing to money in the bank is typically 90 to 120 days. Timelines are typical, not promised — and no offer is guaranteed.
Published deals, and buyer ranges shown at the sizes we qualify — examples of what archives have sold for, not a promise of what yours is worth. Your number only exists as a written offer after the intake call.
Yes. AI companies train their systems on how real organizations operate, and a nonprofit's records — program administration, service delivery, correspondence, accounting — are that kind of history. You license a cleaned, de-identified copy for a set term. Your originals stay yours, and your programs keep running as they do today.
It is removed before any AI developer sees the data. Names, emails, phone numbers and addresses are replaced — a person becomes "Person 01" — dates are generalized, and card numbers and anything about minors are left out entirely. You pick which systems go, and you can see a cleaned sample first. You can also keep beneficiary records out entirely and license only operations and administrative records. Your own counsel makes the final call for your organization.
For most nonprofits, yes. A license covers records the organization built over years, so it belongs in front of the board. The board sees the offer in writing before anything is signed, your counsel can review every document, and whether the license fits the mission is the board's call. If the board says no, nothing has been handed over.
Yes — always. What gets licensed is a copy: you keep your originals, keep using them, and your daily operations don't change. You also pick which systems and which years go; the rest never leaves.
It depends on three things: how many desk staff worked in the systems, how many years of records you hold, and how many sources the work ran through. Case studies from AI labs run from $150K to $550K, and higher for large archives. Your number only exists as a written offer after the intake call — the estimator gives you the rough range first.
Nothing, at any step. Your organization never pays anything.
We never touch or access your data. Not during the intake, not during the export, not ever. Your records go straight from your systems to the AI company you sign with — we are never in that chain. Your business never pays anything.