Sell law firm data to AI companies

What law firm data do AI companies buy?

AI companies buy law firm data that shows how lawyers actually draft, revise and reason. A signed contract on its own is worth less than the same contract with three rounds of tracked changes, the partner's comments and the file memo explaining why a clause was rejected.

Legal AI products are built to draft, review and answer questions the way a trained lawyer would. Training and testing them needs real examples of expert work, which only practicing lawyers produce. Mercor's listing covers firms of 5 to 300 lawyers and staff, so mid-sized and regional practices are in scope.

• Precedents and template banks, with the drafting notes that explain each option
• Contract markups that keep the full redline history between the parties
• Research memos and know-how notes written by associates and partners
• Advice letters, pleadings, briefs and motions, once de-identified
• Matter checklists, file opening procedures and supervision notes

Firm-created know-how is usually the cleanest part of the archive because the firm wrote it and owns it. Client files are more valuable to buyers and much harder to clear, as the sections below explain. See expert domain data for how buyers use this kind of material.

How much do AI companies pay for law firm data?

AI companies publish a range of $50,000 to $1.2 million for law firm data. That is Mercor's referral listing for firms of 5 to 300 lawyers and staff, as reported by ddiy.co in October 2026.

Treat a published range as a ceiling for strong archives, not a quote. AIData.Broker's analysis, reported by ddiy.co, puts the median of buyers' published offers at $280,000 per company but the median of public sales at $31,600. The same analysis finds an exclusive license fetches about 2.5 to 5 times a non-exclusive deal, which is why the exclusivity decision moves the price so much.

What moves a law firm toward the top of a range is volume across many years, documents with revision history, a clear mix of practice areas and a rights position that has been checked. A free data valuation gives you a written range before you commit to anything.

Can a law firm sell client files without breaching confidentiality?

A law firm can only license material that reveals information about a client's matter if the client gives informed consent or a rule permits the disclosure. ABA Model Rule 1.6(a) says a lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent, the disclosure is impliedly authorized to carry out the representation, or paragraph (b) permits it.

The duty does not end with the matter. Model Rule 1.9(c) bars a lawyer from revealing information relating to a former client's representation, or using it to the former client's disadvantage, except as the Rules permit. Rule 1.6(c) also requires reasonable efforts to prevent unauthorized disclosure of, or access to, client information.

The ABA has addressed AI directly. Formal Opinion 512, issued on July 29, 2024, says a client's informed consent is required before information relating to the representation is input into a self-learning generative AI tool, according to the National Conference of Bar Examiners' summary. Licensing client files to an AI company for training raises at least the same question.

Each state adopts its own rules of professional conduct and issues its own ethics opinions. Florida Bar Ethics Opinion 24-1, adopted in January 2024, says lawyers using generative AI must protect the confidentiality of client information. Whether a de-identified file still reveals information relating to a representation is a question for your general counsel or ethics counsel, so take advice under your state's rules.

A law firm should leave out any material where a court order or contract outside the firm's control prevents disclosure, whatever the client consent position. The table sets out the usual categories.

Frankfurt Kurnit Klein and Selz advises excluding high-risk data entirely rather than relying on de-identification, because once a model is trained the data cannot practically be pulled back out. For a law firm, that is a sensible default for criminal defense, government investigation and family law files.

How does de-identification work for legal documents?

De-identification for legal documents means removing far more than client names. A hearing date, a court, an unusual asset and a town can identify a party as reliably as a name.

• Case numbers, court names and hearing dates
• Employer identification numbers, state entity numbers and registered addresses
• Distinctive facts, such as a rare asset, a reported incident or a named property
• Opposing counsel, expert witness and judge names
• Signature blocks, email footers and document metadata

Buyers will ask how you de-identified the files and whether the method was tested. Our vetting standard requires a documented method and a re-identification test on a sample. The guide to de-identification for AI training data covers the methods in detail.

If you hold UK or EU data

Firms with English matters face the SRA's rules as well. Paragraph 6.3 of the SRA Codes of Conduct requires client affairs to stay confidential unless disclosure is required or permitted by law or the client consents, and the SRA's guidance says the duty continues after the client dies.

Two UK statutes add hard exclusions. CPR 31.22 limits the use of documents disclosed in English litigation to the proceedings, and section 333A of the Proceeds of Crime Act 2002 makes it an offense in the regulated sector to disclose a suspicious activity report in a way likely to prejudice an investigation. Exclude both categories.

How does selling law firm data work with Sell My Data To AI?

Selling law firm data with Sell My Data To AI starts with a 30-minute consultation about what the firm holds and where it sits, usually a document management system, email and a practice management system. No records leave your firm at this stage.

If buyers are likely to be interested, we help you build a data manifest and an exclusion list, run our eight vetting checks and show a redacted, randomly drawn sample only to buyers you approve. You sign the license directly with the buyer after your own lawyer has reviewed it, and our success fee is agreed in writing before any introduction.

Book a free consultation to find out whether your archive is likely to interest buyers.