De-Identification: How Personal Data Is Removed Before Anything Moves
De-identification removes or masks personal information — names, email addresses, phone numbers, postal addresses and direct identifiers — from records before they are transferred under a license. What remains is the structure of the work itself, which is what buyers are paying for.
SellMyDataForAI.com helps U.S. companies with 20+ employees license their operating records — email, chat, CRM and project data — to AI labs and training-data buyers. Owners keep ownership of their data; valuation starts with a metadata-only manifest, and no raw data is shared to get an estimate.
What exactly is removed?
Direct identifiers (names, emails, phone numbers, addresses, account numbers) are removed or replaced with consistent placeholders, and free-text is screened for personal details that structured fields miss.
Consistent placeholders matter: “Customer A” stays Customer A across a thread, so the record still shows how a case unfolded — who decided what, in what order, with what result — without identifying anyone.
What is kept, and why?
Kept: the work itself — ticket text with resolutions, decision threads, document structure, timestamps, categories, outcomes. Buyers train models on how work is done, and that signal survives de-identification intact.
This is also why buyers state that personal data is excluded rather than merely handled: the value was never in the identities. A manifest therefore never needs personal data, and neither does the valuation.
Who performs de-identification, and under what standard?
It is a negotiated license term: the buyer, the seller or a specialist performs it, to a standard written into the license, before full delivery. A redacted sample under NDA proves the standard before the full archive moves.
Some records are excluded rather than de-identified — privileged material, patient information, anything a contract keeps out. Exclusion is always an available answer, and the rights check decides which records get which treatment.
How is de-identification verified before delivery?
Verification happens on the redacted sample, before the full archive moves: the buyer confirms the sample meets the license's de-identification standard, and the owner confirms nothing in it identifies a customer or employee. Only then is the full delivery prepared to the same standard.
This sample-first order is deliberate. A standard that looks fine in a contract clause can fail on real records — a name buried in a ticket's free text, an email signature in a forwarded thread. Catching that on a sample of a few hundred records costs an afternoon; catching it after delivery costs the deal's trust. Owners approve the sample themselves, so the check is not delegated to the buyer or to this service: you see exactly what leaves, in the form it leaves, before the license is performed.
Questions owners ask first
Is this a sale or a license?
A license. You keep ownership of your records and keep using them in your business. A buyer pays for the right to use a de-identified copy for AI training, on terms — length, exclusivity, permitted use — that you approve before anything is signed.
Do I have to send my data to get an estimate?
No. The estimate and the first call use only what you describe: systems, years and rough counts. Buyers bid on a metadata-only manifest. Full data moves only after you sign a license directly with a buyer, and personal data is de-identified before transfer.
What does it cost me?
$0 upfront. There is no retainer and no listing fee. We are paid a success fee only if a deal closes — the terms are stated in writing before any submission is made.