---
title: "Sell your company data to AI labs at a price per gigabyte."
description: "Yes, when you hold the rights to the exports. Second Exit licenses an anonymized copy of your threads, tickets, code, documents, photos and videos, and pays you a fixed price on delivery. Your shares, customers and original files stay with you. Operating and closed companies follow the same five steps, from the free estimate to payment."
canonical: "https://laboetie.io/products/second-exit"
language: en
---
# Second Exit: Sell your company data to AI labs at a price per gigabyte.

## Why AI labs buy company data

Language models were trained on what the public internet contains: articles, forums, open source code, books. Labs train the next generation as agents that complete tasks inside a company. For that, a model needs examples of finished tasks, in order, across the people who handled each step. One such example runs from the request and the discussion that clarified it to the ticket, the code change, the review, the release and the reply to the customer.

That sequence is stored in company tools that search engines do not index. AI labs and agent builders price it by scarcity. An acquirer values a company on revenue and growth, so a small startup with modest revenue gets a modest valuation. A lab’s price for the same company’s record depends on how rare its workflows are. For a company with years of linked history, the lab’s figure exceeds the valuation.

## Who pays whom, and where our margin comes from

Three parties take part in each sale: you, La Boétie, and the model builders who license the corpus.

We license the same anonymized corpus to several buyers. We use the income from several licences to pay a price above what a single buyer would pay. We carry the resale risk. Your price, fixed at signature and paid on delivery, does not depend on how many licences we sell afterwards.

## The sample: one workflow across teams

To price your record we ask for one transversal workflow, which is a single task your company carried from A to Z through the teams that touched it. From the sample we assess the continuity, depth and cleanliness of the rest of the record.

## The six steps of the anonymization pipeline

The same pipeline processes the sample and, later, the full record. You receive the anonymized output of your sample before our quote and check it yourself.

- Read only export: We take exports from the tools you already use (Slack, Teams, Google Workspace, Microsoft 365, Jira, Linear, GitHub, GitLab, Notion, Drive, mailboxes, photo and video libraries). Our access to your systems is read only, and we install no software.
- Detection: Every document goes through entity recognition and pattern rules for names, emails, phone numbers, postal addresses, bank and card numbers, national identifiers and customer references. A separate scanner finds credentials: API keys, tokens, passwords, private keys.
- Consistent pseudonyms: Each person and each customer becomes one stable placeholder across every tool, so the thread in Slack, the ticket in Jira and the commit in Git still point to the same actor.
- Media treatment: We blur faces and licence plates. We also strip metadata and GPS coordinates and mask customer data visible on screen recordings. We transcribe speech and scrub the transcript like any other text.
- Human review: A reviewer reads a random share of every batch looking for anything the automated passes missed. A batch with a residual identifier goes back through the pipeline before it moves on.
- Key destruction: The table linking placeholders to real identities is used during processing only and deleted once the batch is accepted. After deletion, no stored table links a placeholder to a person.

## The data perimeter of the sale

- Internal conversations: Slack, Teams, email threads between colleagues.
- Work tracking: tickets, boards, specifications, meeting notes.
- Code and its history, including reviews and incidents, if you choose to include it.
- Documents, spreadsheets and design files with their revisions.
- Photos, screen recordings and videos produced during the work.

## How to sell company data in five steps

The estimate, the sample and the quote are free, and you can stop at any point. You decide whether to sell once you hold a firm price.

- Estimate: Answer the five questions of the estimator below. The estimator shows a range and the multiplier applied for each answer.
- Sample: Under a confidentiality agreement (NDA), you pick one workflow and export it. We anonymize it and send the anonymized version back to you.
- Grading and quote: We grade the sample on continuity, depth, media share and cleanliness, then issue a firm price per gigabyte of anonymized output.
- Extraction: We process the full record in batches. Each batch is measured in gigabytes after anonymization, and you see the running total.
- Payment: We pay you the quoted rate times the delivered volume. Processing, storage and review are at our cost.


## Which company data AI labs want

AI labs train agents to finish tasks at work. An agent needs examples of how a task moves through a company: the request, the debate, the decision, the work, the review, the delivery. A public website holds the result. Your internal tools hold the path.

| Data | Value to a lab | Why |
|---|---|---|
| Threads tied to tickets and commits | High | Shows decisions and their outcome |
| Code with reviews and incidents | High | Shows how experts correct work |
| Screen recordings of real tasks | High | Shows tool use step by step |
| Field photos and repair videos | High for multimodal models | Real tasks are rare in public data |
| Isolated documents | Medium | The result without the process |
| Marketing files | Low | Already resembles public text |
| Customer lists | Excluded | Personal data with no training need here |

Continuity counts for most of the price. Three years of messages with broken links to tickets are worth less than one year where a task can be followed end to end.

## How to prepare your record

You can raise the value of a sale before any sample leaves your company.

1. Check your retention rules. Where your legal duties allow it, pause automatic deletion on chat, mail and ticket tools while you decide.
2. Export while you still can. A closed company loses workspaces when subscriptions end.
3. Find the longest tasks. Pick one that crossed several teams.
4. List your tools and their dates, and note any gaps.
5. Note what sits outside the main tools, such as phone videos or shared drives.
6. Identify files that must never leave, such as medical or legal records.

| Situation | Priority |
|---|---|
| Operating company | Choose the scope, and agree it internally |
| Company closing down | Export before accounts lapse |
| Company already closed | Find who holds the admin rights and the archives |

## The legal questions

A sale of this kind touches several areas of law. The answers depend on your country and your contracts, so have your counsel read the sale agreement.

| Question | Who to ask | What to check |
|---|---|---|
| Who owns the data? | Your counsel | Company ownership, and any liquidator's role if the company closed |
| Do customer contracts forbid this? | Your counsel | Confidentiality and data processing clauses |
| Are employees covered? | Your counsel | Employment terms, and any works council duties |
| Does a third party hold rights? | Your counsel | Licensed content, software and client deliverables |
| Is the data truly anonymous? | Your data protection adviser | The method and the residual risk |

Under the GDPR, Recital 26 places truly anonymous data outside the regulation. Pseudonymous data stays personal data, since someone could link it back. That is why the last step of the pipeline deletes the table linking placeholders to people. Anonymisation is itself a processing of personal data, so it needs a lawful basis too.

Anonymisation cannot erase everything. A rare job title, an unusual event or a writing style can point to a person. Human review looks for those clues, and buyers' contracts forbid reidentification. The risk is reduced and it remains above zero.

## What a sample tells a buyer

| Question | What the review reads |
|---|---|
| Does the task continue across tools? | Links between messages, tickets and code |
| How deep is it? | Number of steps and revisions |
| How clean is it? | Noise, missing files, broken threads |
| Does it include media? | Screens, photos, video |
| Is it rare? | Domain and type of work |

You see the anonymized output of your sample before the quote, and you can check for yourself what would leave.

## Second Exit or a data broker

| Question | Broker on commission | Second Exit |
|---|---|---|
| Who finds the buyer? | The broker, after signature | La Boétie, which carries the resale risk |
| When are you paid? | Often when a buyer pays | On delivery |
| Price | Often an estimate | A firm price per gigabyte after the sample |
| Who does the anonymization? | Varies | La Boétie |
| Margin | A percentage | Stated: licence income minus your price and processing |

A commission broker can earn you more if a large deal closes, and you carry the risk that none does. With a fixed price, you give up that upside and get certainty.

## How to compare offers

1. Is the price firm, or an estimate subject to a buyer?
2. When does money reach your account?
3. Who carries liability for anonymization failures?
4. Can you see the output before you sign?
5. Is the licence exclusive?
6. Do buyers' contracts forbid reidentification?
7. Is the table linking identities destroyed, and when?

## Next step

Use the estimator on this page, then [send a brief](https://laboetie.io/brief) with your company's domain, age and tools. Closed companies should start with the export.


## Frequently asked questions

### Can I sell company data to AI labs?

Yes, when you hold the rights to the exports. Second Exit licenses an anonymized copy of your threads, tickets, code, documents, photos and videos, and pays you a fixed price on delivery. Your shares, customers and original files stay with you. Operating and closed companies follow the same five steps, from the free estimate to payment.

### How can the record be worth more than my company?

Two different buyers price them. An acquirer values your company on revenue, growth and risk. A model builder values your record on scarcity, because continuous examples of real tasks finished across teams are stored inside companies. Epoch AI puts the price of a single training task sold to a lab at $200 to $2,000 (January 2026). A record holds thousands of such tasks, so its price passes the valuation of a young company with modest revenue.

### How much is company data worth to AI labs?

Our reference point is a software startup of under ten people, three to five years old, with a record that can be followed across teams: around $500K. The figure rises with technical and manual domains, more people, more years and more photo and video. The estimator shows each increase. We quote the firm price per gigabyte after the sample review.

### Can I sell our Slack data?

You can sell Slack threads as part of a working record, because model builders pay for finished tasks followed across tools. The sample traces one task from the request through Slack or Teams, the ticket, the code and the release. We take a read only export, give each person and customer one stable placeholder across every tool, and delete secrets before delivery.

### Why price company data per gigabyte?

Both sides can measure a gigabyte. At the sample review we fix a rate from continuity, domain and media share. We measure the volume batch by batch after anonymization, so you are paid for the delivered volume and can check every figure. You see a range in the estimator first, and we set the firm rate after the sample.

### Who buys the data?

AI labs and companies building agents buy the corpus for model training and evaluation on realistic work. Anthropic leaders discussed spending over $1 billion on such training environments within a year (The Information, September 2025). Every licence forbids attempts to reidentify people and any use outside training and evaluation. Buyer names stay confidential under their contracts.

### Is it legal under the GDPR?

The raw export contains personal data, so we process it as your processor under a data processing agreement, as Article 28 of the GDPR (the EU data protection regulation) requires. Pseudonymized data still counts as personal data. Recital 26 places anonymous data outside the regulation, which is why we remove identifiers and delete the mapping table before delivery. The contract lists the excluded data types before the first export.

### Can my competitors read my internal discussions?

Buyers receive the corpus under a licence limited to model training and evaluation. Before delivery, our pipeline replaces your company name, product names, customers and people with placeholders. You choose the perimeter and can exclude any project, client or piece of code from it. Their contract forbids any attempt to reidentify a person or a company.

### My company shut down. Can I still sell?

Yes, as long as the exports still exist and the seller holds the rights to them. After a closure, the liquidator or the former directors hold those rights. The process stays the same. One workflow goes through the sample review. We then quote a firm price per gigabyte of anonymized output.

### Do I lose the use of my own data?

No. The contract grants a licence to an anonymized copy of the record. Your originals stay on your systems. You keep your rights to use them. We license the anonymized corpus to several buyers, and your price stays fixed at signature whatever number of licences follows.

### When do I get paid?

On delivery we pay the quoted rate per gigabyte times the delivered volume. You see the running total as we measure each batch after anonymization. The price is fixed at signature, we carry the resale risk, and processing, storage and review stay at our cost. The estimate, the sample and the quote are free, so you can stop at any point before the sale.
