Company data licensing

Sell your company data to AI companies

AI labs pay to license real business records as AI training data: Slack and email threads, CRM history, support tickets and internal docs. Answer a few questions and we'll tell you whether yours could qualify. It's a licence, not a sale, so the data stays yours.

Check your data
See what labs buy

Where the data lives

The data AI labs pay for is already in your tools.

Slack threads, email, support tickets and CRM history record how your business actually makes decisions. Years of those records are what labs license for training. If your team has used tools like these for a few years, it's worth checking what you hold.

Slack Microsoft Teams Gmail Outlook Zoom Google Meet Google Drive SharePoint Notion
Confluence Dropbox HubSpot Salesforce Zendesk Intercom Jira Linear Asana
ClickUp Trello GitHub GitLab Figma QuickBooks Xero SAP Shopify

Data licensing, explained

What is data licensing?

Data licensing is when a company gives an AI company the right to use copies of its business records, such as chat threads, support tickets, CRM history and documents, as AI training data for a set purpose and period. The company keeps ownership, and a non‑exclusive licence lets the same records earn more than once.

Paid by the companies building AI

AI labs and the data companies that supply them pay to license real business records.

Licensed, never sold

You grant the use of a copy for a set purpose and period. The original stays yours.

Only what you own

Only records your company has the rights to are licensed. Data you hold for clients stays out of it.

Answer a few questions and we'll tell you whether yours could qualify.

Check your data

Why now

Why AI companies are buying business data now

AI models have already trained on most of the public internet. What they still lack is a record of real work: decisions made across weeks of messages, approvals and fixes. That record only exists inside companies, so AI companies are now paying to license it.

One AI data company

$1B
committed in October 2026 to license company data within a year

One airline's records

$10M
Google's winning bid for a failed airline's emails and Teams chats, 2026

One healthcare dataset

$200M
in data licensing and model fees for one oncology dataset, 2025

What labs actually buy

What AI companies look for in business data

Size alone doesn't decide what records are worth. AI labs pay for depth: records that show how a business actually makes decisions, and what happened next.

Years of history

A long-running record shows how decisions played out over time, not just a single moment.

Outcomes attached

Closed deals, resolved tickets and paid invoices let a model connect a decision to its result.

Reasoning written down

Threads where people explain why they chose something, not only what they did.

Specialist work

Decisions from finance, legal and engineering teams are among the records buyers name most often.

What qualifies

What business data qualifies?

Buyers mostly look for companies with 30 or more people, records in English and at least two years of history, with US companies first in line. If that sounds like you and your company owns the records, it's likely to count. The free check confirms it.

Conversations

Chat and email threads where real decisions get made.

SlackMicrosoft TeamsGmailOutlook

Calls and meetings

Recordings and transcripts of teams talking problems through.

ZoomGoogle Meet

Documents and knowledge

SOPs, proposals, reports and wikis: how the company says work should be done.

Google DriveSharePointNotionConfluenceDropbox

Sales and CRM

Deals from first touch to closed or lost, and every step in between.

HubSpotSalesforce

Customer support

Problems captured end to end, from complaint to resolution.

ZendeskIntercom

Projects and engineering

Plans, issues, designs and project history, from idea to shipped.

JiraLinearAsanaClickUp+4

Finance and operations

Invoices, reconciliations, approvals and orders from the systems that run the business.

QuickBooksXeroSAPShopify

Not sure your stack counts?

Tell us what you use and we'll give you a straight answer, even if it's no.

Check your data

Out of scope: data you hold on behalf of clients, and anything the company doesn't have the rights to. Got private codebases? License your code separately →

Who it's for

Which businesses can sell their data to AI companies?

Sector matters less than depth. If your team has years of records showing how work gets done, you're likely a fit.

  • Software and SaaS

    Support queues, product specs and sprint history.

  • Agencies and consulting

    Proposals, internal project threads and retros.

  • Finance and accounting

    Month-end close, reconciliations and approvals.

  • Legal and compliance

    Contract reviews, redlines and compliance calls.

  • Healthcare

    Scheduling, billing, claims and admin workflows.

  • Customer support and BPO

    Tickets, chats and escalations with outcomes.

  • E-commerce and retail

    Orders, returns and customer conversations.

  • Logistics and supply chain

    Dispatch, exceptions and carrier escalations.

  • Recruiting and staffing

    Pipelines, screening notes and placements.

  • Construction and trades

    Job tickets, site reports and change orders.

  • Gaming studios

    Design docs, build histories and playtest notes.

A free check tells you whether your records could qualify.

Check your data
What qualifies

How it works

How to sell your company data to AI companies

Four steps, and nothing is licensed until you've agreed the terms.

  1. Check

    01

    Tell us which tools your team uses and how many years of history they hold.

  2. Review

    02

    We look at what you have and tell you whether it could qualify and what affects its value.

  3. Agree terms

    03

    You decide which sources, date ranges and uses are included before you sign.

  4. Get paid

    04

    The data is licensed for a set use and term. You're paid for the licence and keep the original.

Free to check, no commitment.

Check your data
Licensing vs selling

Licence or sale

Licensing your data vs selling it

A sale hands the data over for one price. A licence lets the buyer use a copy on agreed terms, and you keep the original.

Question
A licence
A sale
Who owns the data after
You do
The buyer
Can you keep using it
Yes, as before
Depends on the deal
How the buyer can use it
Only for the agreed use and term
Usually without limits
Can it earn again
It can, if the terms allow
No, it's a one‑off
How you're paid
A licence fee
A one‑off price

Key terms

The terms, in one place.

What a deal with us looks like, before you fill anything in.

Who pays you
AI labs and the data companies that supply them. Pangea runs the check and handles the process.
Deal size
Around $100K to $2M, and more for large, deep records. You get a real figure after a review.
Structure
A licence, not a sale. The buyer uses a copy for an agreed purpose and term.
Cost to you
Nothing. No fee to check and nothing to pay upfront.
Who qualifies
Any sector. What counts is years of records showing how work gets done.
Scope
You approve it before you sign: which tools, which date ranges and which uses.
Personal details
Removed first. Names, emails, phone and account numbers come out before the record is shared.
Access
Read-only, and only to the sources you approve. Nothing is changed or deleted.
Ownership
You keep the original and keep using it as before. A non-exclusive licence can earn again.

Free to check, no commitment.

Check your data
Read the FAQ

Frequently asked questions

Everything you ask us first.

Still unsure? Start the check. Someone reads it and replies, even if the answer is that your data isn't a fit.

Check your data

How much is my company's data worth?

It depends far more on depth than on size. A record covering many years, with outcomes attached (closed deals, resolved tickets, paid invoices) and specialist decisions in it, is worth much more than a large pile of shallow data. Company data licences in the market currently range from around 100K USD to over 2M USD for large, deep records. You only get a real number after a review.

Start with a check of what you hold: which tools your team uses and how many years of history they contain. If there's a fit, you agree the scope and terms, personal details are removed from a read-only copy, and the de-identified record is licensed to an AI lab for a set use and term. You're paid for the licence and keep the original data.

Generally yes, when the company owns the data and personal information is handled lawfully. Three things decide it: whether the company owns the records, what your customer and supplier contracts allow, and how personal data is removed. All three are checked before anything is licensed. This isn't legal advice, so involve your own counsel before you sign.

AI labs building frontier models, and the data companies that supply them. They want real records of how work gets done, because that is the part of the world models haven't been able to learn from public text.

With a licence you keep ownership and the buyer may only use the data for a set purpose and term. With a sale, ownership moves to the buyer. Licensing means you keep using the data exactly as before, and the record can sometimes be licensed again later.

No. Names, email addresses, phone and account numbers and addresses are removed or replaced before the record leaves processing. Roles, timing and the work itself are kept, because that's what carries the value.

No. The first check only asks which tools you use and roughly how much history they hold. If you go ahead, access is read-only to the sources you approve, and extraction and de-identification are handled for you.

Nothing. There is no fee to check your data and nothing to pay upfront.

The check is a short form. A person reviews every submission and gets back to you. From there, timing depends on how much data you have and what buyers are looking for, so we'll give you a realistic timeline once we've reviewed yours.

No. Only records the company owns, such as its own chat, email, CRM and files. Anything you hold on behalf of clients is out of scope.

They can be licensed on their own, at standing per-token rates, separately from the rest of your data. See the code licensing page for how that works.