# How much is law firm data worth to AI companies?

> Legal AI tools need real legal work to learn from, and most of it sits inside law firms. Here is what makes a legal archive valuable to AI buyers, what lowers its value and the price ranges buyers have published so far.

How much is law firm data worth? Until recently, almost nothing outside the firm. That has changed. AI companies are building tools that draft, review and research like lawyers do, and they need real legal work to train them. This guide is the legal version of our general guide on [how much company data is worth](https://corpusmarket.com/guides/how-much-is-company-data-worth.md). It covers why buyers want legal data, which document types count most, what raises and lowers value, and the price ranges that have been published.

## Why AI companies want legal data for AI training

Language models learned to write from billions of public web pages, and they learned a fair amount of law from published court opinions and statutes. What they haven't seen much of is the work behind those documents – the drafts, the research, the negotiation and the strategy that never get filed or published.

Demand is coming from several directions:

- **Legal AI assistants** for contract drafting and review, legal research, due diligence and litigation support, sold to firms and in-house teams.
- **AI labs training agents** meant to carry out longer professional tasks, where legal work is one of the clearest test cases.
- **Buyer programs** that publish ranges for law firm data and ask for closed matters with complete records from several years of practice.

The upshot: a document management system full of closed matters may now be worth assessing. Through a license, the firm keeps ownership and chooses what is included.

## What public legal datasets lack

To see why your archive could matter, look at what researchers train on today:

- **Court opinions.** The Caselaw Access Project made millions of published US decisions freely available. Opinions show how courts ruled, not how the lawyers prepared the case.
- **Public contracts.** Many legal datasets, such as CUAD, are built from contracts filed with the SEC. Those are final, executed versions, without drafts, issues lists or markups.
- **Mixed public collections.** Pile of Law gathers court records, regulations, contracts and other public legal text. It is large, but everything in it was public to begin with.

None of these shows how a matter is actually run. There are no internal memos, no drafts with tracked changes, no research trail behind a motion and rarely a complete record from intake to outcome. That is the gap a law firm's archive fills.

## Value tiers by document type

The key question is how much of the lawyer's reasoning survives in the material.

| Document type | Examples | What buyers get | Relative value |
|---|---|---|---|
| Complete matter files | pleadings, motions, briefs, orders, correspondence, outcome | A full matter from start to finish | Highest |
| Contract drafts with redlines | draft sequences, markups, issues lists, closing sets | How a document was negotiated | High |
| Precedent and form banks | templates, clause libraries, checklists | Curated, reusable legal drafting | Medium to high |
| Research memos and KM content | practice notes, internal memos, training material | Legal reasoning on recurring questions | Medium, higher when linked to matters |
| Filed briefs and motions only | documents from the docket | Final public versions | Medium to low |
| Scanned or image-only PDFs | old paper files, scanned sets | Text that has to be extracted first | Low |

Format matters too. Native Word files with tracked changes and version history in the document management system keep far more than flattened PDFs. The practical advice is simple: don't convert or clean up anything before an assessment. Keep native files and versions as they are.

**[See what your data is worth](https://corpusmarket.com/submit-data)** – Free estimate in two minutes – not an offer

## What multiplies the value

Document type sets the baseline. Context multiplies it:

- **Linked matters.** A motion that sits with its research memo, the opposing brief and the order that decided it is worth more than any of those on its own. Buyers want to connect a question to the work and to the result.
- **Version history.** Drafts and redlines kept in sequence show reasoning that a final document never does.
- **Outcomes.** Knowing how a matter ended – the ruling, the settlement range, the closing – turns a pile of documents into a record a model can learn from.
- **Years and variety.** Several years of practice and more than one practice area give a buyer more to learn from than many near-identical matters. Buyer programs typically ask for five or more years of closed matters.
- **Clean rights.** The buyer needs to know the material can be licensed. Firm-owned work and matters with client consent are worth far more than material with open questions – see [can law firms license client files for AI](https://corpusmarket.com/guides/can-law-firms-license-client-files-for-ai.md).

## What lowers the value

Some things pull value down, or rule material out entirely:

- **Material owned by other parties.** Documents produced in discovery, data room contents and anything under a protective order or seal usually can't be licensed – see [preparing legal documents for AI licensing](https://corpusmarket.com/guides/preparing-legal-documents-for-ai-licensing.md).
- **Client information without consent.** Where client consent can't be obtained, those matters are normally left out, however valuable they look.
- **Heavy personal data.** Matters full of health, financial or other sensitive personal information need far more clean-up, and some are left out.
- **Duplicates.** Copies saved in several workspaces, email attachments and near-identical versions inflate the count without adding information.
- **Only finals or only PDFs.** An archive where drafts were deleted or everything was printed to PDF sits lower in the tiers.

## The price ranges published so far

There is no public price list for legal data, and no one has published a per-document or per-GB price. Anyone quoting a figure per document or per gigabyte is guessing. What does exist is a buyer range per firm.

| Firm type | Published range per firm | What buyers count |
|---|---|---|
| Law firm | about $50,000 – $1.25 million | closed matters with 500+ documents, from 5+ years of practice |

Read these figures carefully:

- **They are ranges, not quotes.** They are what buyers say they may pay, a best case rather than realized prices.
- **They are per firm or engagement.** They are not per document, per matter or per gigabyte.
- **They are not guaranteed.** Every archive and every deal is unique. Your archive can land below, above or outside this range, and some archives can't be licensed at all.

A license also shapes the economics. By default it is non-exclusive and limited in scope, time and use, so the same archive can in principle be licensed to more than one buyer over time. An exclusive license may pay more per deal but closes the door to others.

To see where your archive might land, describe it in our form on [what is your data worth](https://corpusmarket.com/what-is-your-data-worth.md). You get an indicative estimate per buyer, built on published buyer ranges. It is an estimate, not an offer.

## Is our legal data valuable? A quick self-check

You don't need to send any documents to get an assessment. Go through the questions below. The more you can answer yes, the stronger your position:

1. **Closed matters?** Do you have at least a few dozen closed matters with a substantial document count each?
2. **Complete records?** Are pleadings, research, correspondence and the outcome kept together for each matter?
3. **Drafts and versions?** Does your document management system keep drafts, tracked changes and version history?
4. **Firm-owned material?** Do you have templates, clause libraries, practice notes or training material written in-house?
5. **Years and variety?** How many years and practice areas does the archive cover?
6. **Engagement terms?** Do your engagement letters say anything about how the firm may use matter documents?
7. **Client consent?** Are there clients who might agree to their anonymized matters being included?
8. **Third-party material?** How much of the archive is discovery productions, data rooms or documents under protective order?
9. **Personal data?** Are there practice areas with a lot of health, financial or other sensitive personal information?
10. **Order?** Can you tell which documents belong to which matter, or are files scattered with duplicates?

If you are unsure about several of these, that is normal. Working them out is part of the assessment.

## How a license works

You license the data. You don't sell it outright. The firm keeps ownership of its documents, the license is non-exclusive by default, and you approve every deal before anything happens. Buyers only see an anonymous profile until you say yes to sharing more, and your full archive stays in your systems until you have approved a deal. The assessment is free, and we are paid only when a deal goes through.

When you are ready, read how to [sell law firm data to AI companies](https://corpusmarket.com/sell-law-firm-data.md), or start by describing what you have.

## Summary

Legal data is in demand because legal AI tools and agents need to learn from real legal work, and public datasets mostly hold court opinions, executed contracts and other documents that were public to begin with. Complete matter files and contract drafts with redlines sit at the top of the value tiers, and linked records, version history, outcomes and years of practice multiply that value. Material owned by other parties, client information without consent, heavy personal data and duplicates pull it down. Published buyer ranges run from about $50,000 to $1.25 million per firm. They are ranges, not quotes, and no per-document price is public. If your archive passes the self-check, it may be time to [license your firm's legal archive](https://corpusmarket.com/sell-law-firm-data.md).

## Frequently asked questions

### Is our law firm's data valuable?

It can be, if you have complete closed matters, contract drafts with their redlines, or a well-kept precedent bank, covering several years and practice areas. Clear rights matter just as much. Archives made up mostly of client material you can't get consent for, documents produced by other parties or loose PDFs are worth far less.

### What do AI companies pay for legal data?

There are no public per-document or per-GB prices. Published buyer ranges for law firms run from about $50,000 to $1.25 million per firm. These are ranges, not quotes, and many archives land below them or can't be licensed at all.

### Are drafts and redlines worth more than final documents?

Usually, yes. A final contract or filed brief shows the result, while the drafts and markups show how lawyers got there – which issues were negotiated, what was given up and why. That sequence is hard to find anywhere public, so it is what many buyers ask for.

### Are public court filings worth anything?

Some, but less on their own. Filed briefs and motions can be found on court dockets, so buyers value them most when they come with what isn't public – the research, the drafts, the opposing papers and the outcome of the same matter.