Public sites show what people say. Company records capture decisions, exceptions and outcomes.
Get paid for data your company already has.
AI labs license records of how work gets done.
- RecordsCompany email, documents and CRM history
- PrivacyIdentifying details removed under agreed terms
How it works.
Four steps from first estimate to payment. Nothing leaves your systems until you’ve signed off on the scope.
- 1
Get started
See what your data could be worth in two minutes, then answer four short steps about your systems.
Run the estimate - 2
Review
We check your answers and tell you whether your data fits and which route suits your company’s size, usually within a few working days.
- 3
Discovery
If there’s a fit, we ask a few follow-up questions and book a call about your data, your privacy needs and what you want to license.
- 4
Agreement and payment
You approve the terms in writing. The data is de-identified, a lab licenses it, and you’re paid in as little as 7 days.
Agreed with you before anything moves
- What’s in scope, and what stays off limits
- Security and confidentiality standards
- How identifying details are removed or replaced
- Who at your company signs off
$100K+A qualified partnership: a mature company with well-documented work.
$500K+Large-scale datasets, or ongoing supply across several teams and functions.
$1M+Highly unique operational data that few other companies hold.
Pay is set deal by deal on dataset size, how complex the workflows are, data quality, the expertise captured and how rare the knowledge is. Estimate yours.
Instant estimate · about 2 minutes
What could your data be worth?
Build a quick range from your company profile and the work systems you use. Nothing is saved.
01
Company profile
Start with team size and how far your records go back.
02
Work systems
Select every system that holds useful work history.
Choose all that apply
Multiple selections are fine
Do you qualify?
These are the buyer requirements AI labs check before they license a company’s data. The more of them you meet, the stronger your estimate holds up in review.
Size and history
At least 25 employees. For US companies that means W-2 employees; contractors don’t count. 25–50 employees: direct route. 50–100+ employees: 6/8 Labs review. Several years of records help establish a useful history.
Direct or reviewedDocumented workflows
Support resolutions, project threads, SOPs, decision records and investigation histories.
Problem → decision → outcomeVolume and structure
Enough history that the cleaned dataset is still substantial. Sparse CRM fields, marketing content on its own and thin logs usually fall short.
Depth over sizeEnglish, in countable systems
Mostly English-language records in exportable tools such as Slack, Teams, email, Zendesk, Salesforce or Notion.
Exportable toolsClean rights
The company must own or control the right to license these records. Contracts or third-party rights can limit use.
Yours to licenseDe-identifiable
Identifiers must be removable without breaking the record sequence. Patient, clinical and genomic data often cannot meet this requirement.
Redactable
Industries that fit
- TechSoftware, SaaS and engineering teams.
- ScienceBiotech, pharma, labs, chemicals and manufacturing operations.
- ServicesConsulting, professional services, BPOs and recruiting firms.
- FinanceFinancial services, accounting, reconciliations and reporting.
- LegalContracts, policy, compliance and governance.
- OperationsLogistics, fulfilment, inventory and field services.
- SupportTicket workflows, QA processes, customer operations and knowledge bases.
- SalesCRM data, sales playbooks, customer lifecycle and revenue operations.
Where demand is strongest
- LocationUS companies first, then the UK, Canada and Australia. We consider companies anywhere.
- LanguageDocumentation and day-to-day work mainly in English.
- ToolsActive use of modern software, so the work is on record.
- Not sure?Check your data and we’ll tell you where it fits. It takes about three minutes and commits you to nothing.
The market, in numbers.
Figures published by the companies building this market. Each belongs to the company named above it.
Maryah
$20M+Paid to companies
For licensed business data.
100+Active buyers
Looking for business data now.
4×Sales per dataset
On average, one dataset is licensed to several buyers.
Grepped
$5MTop estimate per company
Its range for one company’s data starts at $20K.
<5%Real jobs AI can finish
The reason labs pay for real examples of work.
7 daysFastest payout
After a buyer purchases.
$100KTop referral fee
10% of the deal, paid to whoever refers it.
Why business data is worth more every year.
One AI data company’s announced 12-month commitment to acquire and license enterprise operational data. Planned spend, not spend already made.
Completeness, record quality and years of useful history. See what can qualify or check your data.
Read the licensing guide
The public web
Already trained on
Mostly public writing about work.
Inside your company
What labs pay for
Records of how work gets done.
What we do
What we look for
Companies with continuous, well-recorded business data and documented ways of working.
What you earn
If an AI lab buys your listing, payment is due within 60 days. Some deals pay in as little as 7.
Four questions to answer before a valuation.
A clear inventory and a careful read of the terms make the process easier to assess.
01
Which systems hold the records, how far back they go and how much can be exported?
02
How complete is the history, and does it capture the steps and decisions behind the outcome?
03
Did a change in messaging tools leave gaps in the team’s conversation history?
04
What does the licence allow, and who is responsible for handling and deleting the data?
We work through these questions with you before a listing reaches AI labs.
The value is in how the work got done.
AI labs look for the request, the steps, key decisions, checks and result. You may find that history in the tools your teams already use.
Messages
Team chat and email where work gets coordinated, including the decisions and the reasoning behind them.
Slack, Teams, Gmail, OutlookDocuments
Procedures, playbooks, knowledge bases, contracts and reports: how the company says the work should be done.
Google Drive, SharePoint, Notion, DocuSignRecords and raw data
Pipelines, deal histories, invoices, ledgers and orders from the systems that run the business.
Salesforce, HubSpot, QuickBooks, XeroTasks
Work captured end to end: the brief, the back-and-forth and the finished result.
Zendesk, Jira, Asana, ServiceNowCodebases
Repositories with their history: the commits, reviews and issues that explain why the code is the way it is.
GitHub, GitLab, BitbucketExpert judgment and feedback
Records of how teams weigh information, make decisions and assess AI output in their work.
Across every system
What buyers want most
Problem decision outcome.
Labs look for records that connect a problem, a decision and its outcome.
Names and other identifying details are removed or redacted before a buyer sees the data, under the terms agreed with the company.
- Lab and process procedures and protocols
- Project histories showing how experiments were designed, run, failed and iterated
- Deviation and quality investigations, including corrective actions
- Tech transfer packages and manufacturing process documentation
- Decision threads in Slack, email or Teams on experiment design, resourcing or process changes, after strong de-identification
- Knowledge bases, checklists and handover documents used by scientific and operations teams
Start with a data inventory.
List the systems, date ranges and record counts before a valuation. Buyers need those details to assess a dataset.
01
Every workspace
Every system that holds business records: team chat, email, CRM, support and documents.
02
Years available
How far back each system can export. That’s often shorter than the company’s history.
03
Total quantity
TBs, messages, tickets or deals. Whatever unit fits the system.
04
Ready for review
A clear inventory helps us decide what to review next.
Know what it’s worth before you list.
An estimate can help set expectations. A buyer’s offer depends on review of the records and the licence terms.
Some labs
Some estimates focus on headcount and storage volume.
How we value it
We also consider record quality, workflow detail and continuity.
- Volume and record quality
- Workflow detail and expertise captured
- How complete and continuous the history is
A company may have older records even after changing systems. The timeline below shows how a tool change can break conversation history.
Email
Gmail
Documents
Google Drive
CRM
HubSpot
Internal messaging
WhatsAppDiscordSlack
Full-context credit
2 years
From inventory to de-identified.
The period between extraction and de-identification needs clear safeguards. Put the handling, timeline and responsibility in the agreement.
01
Inventory
Your systems, years and volume, listed before the first call.
02
Agreement
Price, exclusivity, acceptance, payout and scope of use, agreed and signed.
03
Extraction
Raw files leave your systems. Exposure starts here.
04
De-identification
We ask for it within 14 days, scaled to data size, done by an independent third-party firm.
Before
Person A approved the revised delivery plan for Company B. Send the signed copy to person@example.com and refund the card on file.
After de-identification
Person 14 approved the revised delivery plan for Company 08. Send the signed copy to person14@example.com and refund payment detail removed.
Ask for a Certificate of Deletion after the raw files are transformed, and set the timing in the agreement.
Refer a company. Earn 8.5%+, up to $100K.
- 6/8 vetting8.5% of the deal or 12% of what the company is paid, whichever is higher, up to $100K per company, for companies with 50 to 100+ employees. Paid on top of the company’s price, within 30 days of us getting paid.
- ProtectedYour referral is protected for 13 months from the day you register it.
- ExampleA $300,000 deal through our vetting pays you at least $25,500, on top of the company’s $300,000.
Company gets$300,000
Your 8.5%+$25.5K+
How our rates compare.
Refer a company through 6/8 Labs and compare what you earn with the published terms of other AI-data referral programs.
| Typical AI-data referral programs | 6/8 Labs | |
|---|---|---|
| Referral rate | 6% to 10%, or a flat $50K to $75K fee | 8.5% of the deal or 12% of what the company is paid, whichever is higher |
| Cap | Often $25K to $100K per company | $100K per company for referrers; none for 50/50 partners |
| Repeat licences | Usually paid once | Paid again on every licence of the same company’s data, for 3 years |
| Recruiter bonus | 20% of what your recruits earn | 20% of what your recruits earn, paid on top of theirs |
| Margin-share partners | 50% of the margin, by application | 50% of the margin, never less than the standard rate |
| Protection | First link wins, or up to 12 months for partners | 13 months for everyone, then the life of the deal once the company starts |
Typical terms come from referral programs published by AI-data brokers and buyers in November 2026, shown without names. The bars apply each rate to what the company is paid; for 6/8 Labs that is our minimum: referrers are capped at $100K per company, and 50/50 partners, who also help close the deal, have no cap and can earn more from half our margin. Examples only, not offers.
Wholesale data
Large archives get a custom deal.
Large archives, multiple businesses or ongoing data supply may suit a custom deal. Terms depend on the collection and buyer.
- VolumeMulti-terabyte archives with years of history across many systems.
- OngoingRecurring supply as new records are created.
- GroupsSeveral companies or brands licensed as one package.
Or write to dhindsamax@gmail.com with a line on what you have. We reply within a few working days.
Know the terms before you sign.
Check what the agreement allows, how long the licence lasts, and who is responsible for handling and deleting the files.
Yes. Labs license records such as email, chat, documents, CRM history and support tickets for AI development.
- The agreement defines permitted uses and term.
- Your company keeps ownership; identifying details are removed under terms you approve.
Value depends on years of continuous history, record quality and the decisions captured.
- Approved packages commonly range from $100K to $2M+.
- Use the instant estimate or full estimate for guidance; neither is an offer.
See the full company data pricing guide.
Company records capture decisions and outcomes that public websites rarely show.
On October 9, 2026, one AI data company announced a $1 billion commitment to acquire and license enterprise operational data.
It covers the next 12 months; this is planned spend, not money already spent.
A broker connects companies with buyers and helps prepare a system inventory, valuation and contract review.
Refer through 6/8 Labs to earn 8.5% of deal value or 12% of what the company is paid, whichever is higher, up to $100K per company. The fee is paid on top of the company’s proceeds.
No. A data licence grants defined uses of the data. You give up no equity and take on no debt, and you keep ownership of the underlying records.
Exclusivity sets how long a buyer has sole use of the data. Check the trade-offs before agreeing:
- How long the exclusive period lasts.
- What the buyer can do during that period.
- How exclusivity affects the price and your ability to license elsewhere.
A 30-day review lets the buyer confirm the files match the agreed inventory. Define review criteria and any right to cancel in the contract.
For deals through our vetting, payout is due within 60 days after purchase; some deals pay in as little as 7 days.
Other deals follow their own contract. Confirm the payment date and conditions before signing.
Check what the buyer may do with the data, for how long, and whether it may be shared further. The agreement should define each permitted use.
Whether sublicensing is allowed, and to whom, depends on the agreement. Name permitted recipients and set limits on further sharing if needed.
Put these details in the agreement:
- Who handles raw files and when de-identification is complete.
- Who is responsible at each stage.
- How and when deletion is documented.
The cap, and what it covers. Check both before you sign.
Payment depends on the company’s referral route:
- 50 to 100+ employees: 8.5% of deal value or 12% of what the company is paid, whichever is higher, up to $100K per company. Paid on top of the company’s proceeds within 30 days after 6/8 Labs receives payment. A $300,000 deal pays at least $25,500.
- 25 to 50 employees: partner rewards up to $50K as the company reaches revenue milestones.
On our deals, identifying information is removed before any lab uses the data. Access is limited to the few people working on your deal, your data is processed on its own and never mixed with other companies’ data, and it’s kept only for the period the agreement sets. You can review a representative sample before it’s used, and deletion requests are honored under the agreement’s terms.
Before signing, confirm:
- Which files are included and who can access them.
- Security and de-identification requirements.
- When originals must be deleted and how use is limited.
- Whether you can review a sample before the data is used.
The company does. A licence grants agreed uses for an agreed period; ownership of the underlying data stays with you.
Payment set by the value of the data, and you keep ownership. Qualified partnerships commonly start at $100K+, large-scale or ongoing data across several teams reaches $500K+, and highly unique operational data can pass $1M.
Each deal is valued on its own: dataset size and volume, how complex the workflows are, data quality, the domain expertise captured and how unique the knowledge is.
It’s processed only for the agreed use. Identifying details are removed, and some datasets are rewritten with consistent synthetic identities that keep the workflow intact. It’s kept for the agreed retention period, you can review a representative sample before it’s used, and it’s deleted on request or at the end of the engagement under the agreed terms.
Companies with 25 to 50 employees use the direct route: a short AI interview to check fit, then partnership discussions, the agreement and onboarding. Companies with 50 to 100+ go through 6/8 Labs’ review. Next steps depend on fit.
Yes. Use the partner link for companies with 25 to 50 employees:
- Get your link from the Partner tab in the referral portal.
- Send it before the company applies; earlier applications can’t be credited later.
- Rewards can reach $50K as the company meets revenue milestones.
We accept referrals worldwide. Current demand is strongest in the US, UK and Canada, with records mainly in English. Your team should also be able to work with us in English through onboarding.
The licence should state whether a buyer may use records to reproduce your products, customer relationships or operations.
Large archives, multiple brands or ongoing data supply may suit a custom deal. Terms depend on the buyer.
Tell us what you have at dhindsamax@gmail.com.
Key referral terms:
- Protected for 13 months after registration.
- Once a data check starts, protection lasts for the life of the deal.
- Commission is 8.5% of the deal or 12% of what the company is paid, whichever is higher, up to $100K per company, paid within 30 days after 6/8 Labs receives payment.
Labs look for records that connect a problem, a decision and an outcome. Examples include:
- Support tickets with full resolutions.
- Project and experiment histories.
- Procedures, investigations and corrective actions.
- Decision threads in email or chat.
Identifying details are removed or redacted under agreed terms.
Check your data.
Answer four short steps. 6/8 Labs reviews your responses and routes you by company size.
Not sure you qualify? Check the buyer requirements first.
What the optional admin check reads
- Aggregate account, email, mailbox and storage counts from Google Workspace or Microsoft 365 usage reports.
- No message or file contents. An administrator must approve the read-only access.
Where you go
It depends on size.
Answer four steps and the check sends you to the right place.
Contact
Questions first?
Ask about terms, pricing or your data check. Larger companies can start review here; duplicate checks protect referral credit.
- 1 · ListTell us the company, its systems, the years available and the volume.
- 2 · ReviewWe check the data and package it, with identifying details removed.
- 3 · PreviewAI labs see redacted previews only. Your data stays under your control until a sale.
- 4 · Get paidA lab licenses the data and you’re paid in 7 to 60 days.
Not your company? Refer it instead and earn at least 8.5% of the deal, up to $100K. Choose “Refer a company” below.
General information only. Ask your attorney to review the Data Licensing Agreement and do your own due diligence.