DATAVALUATION Check eligibility

A private-first qualification screen

Your operations may be worth more than the work they support.

Established companies can be compensated for carefully prepared operational data that helps develop more capable, responsible AI. Find out whether your company is a fit—before you share a thing.

Check your fit How data is protected

Initial qualification only. This page does not ask for files, exports, or customer data.

THE INITIAL SCREEN

04signals of operational maturity
No transfer. No commitment. Just clarity.

01 — ELIGIBILITY

First, check the operating signals.

If each statement is true, your company may be a strong candidate for an initial review. This is a fit screen—not a request to send data.

Operational data eligibility self-check

02 — SYSTEM FOOTPRINT

What counts as operational data?

It is the trace of how work moves through a business: the collaboration, handoffs, records, and outcomes created as teams serve customers and make decisions. Value usually comes from context across systems—not from any single file.

01Workspace

Google Workspace
Microsoft 365

02Communication

Slack
Microsoft Teams

03Customer context

CRM
Help desk

04Execution

Project management
Workflow tools

05Business record

Accounting
ERP

↗ More systems do not automatically mean a better fit. The important question is whether they capture a real operating story with useful context and outcomes.

03 — PROTECTION

A valuable dataset should never be a raw data dump.

Any viable opportunity begins with rights, privacy, and commercial safeguards. A careful process is designed to reduce exposure before a model developer ever uses a dataset.

  1. 01
    Scope before access

    Review what the company can legally share and select only the data relevant to a defined use—not every system record.

  2. 02
    Remove direct identifiers

    Names, emails, account IDs, credentials, and unnecessary personal information are excluded or redacted.

  3. 03
    Protect the business itself

    Trade secrets, pricing, customer identities, code, security details, and rare identifying events are evaluated for removal or transformation.

  4. 04
    De-identify and test

    Data is reviewed for re-identification risk; in higher-risk cases, sharing can be limited to protected environments, aggregates, or synthetic derivatives.

  5. 05
    Control the use

    Purpose limits, security controls, retention terms, deletion obligations, and no-redistribution provisions are part of a responsible agreement.

THE FIRST CONVERSATION

No uploads. No exports. No customer lists.

The initial intake is for high-level company and systems information only. A qualified next step is an evaluation conversation, where rights, exclusions, security, and permitted use are defined before any material is considered.

04 — CLARITY

A few common questions.

Do we need 35 employees today?

No. The screen uses peak headcount. What matters is that the company built meaningful operating history and retained useful records.

Do all of our systems need to be connected?

No. The systems can be separate. A later evaluation considers whether their combined history reflects real workflows, decisions, and outcomes.

What if some communication is not in English?

The relevant stored internal communication should be at least 80% English. Multilingual businesses can still be a fit when the qualifying material meets that threshold.

Is a qualified screen a commitment to share data?

No. Qualification opens an initial conversation. Any data evaluation or license depends on rights, safeguards, scope, and a separate agreement.

05 — DATA INVENTORY

How to inventory your company's data

AI companies license real business records — emails, chats, support tickets, job files, drawings, code — from established companies to train and test their models. We find the right buyer, structure the license and run the process for you. You license the data; you do not sell your company or give up your records.

This form tells us whether your company's data is a fit and roughly how valuable it could be. Filling it in commits you to nothing.

WHAT MAKES A DATASET VALUABLE

  1. People × years

    The more people who worked in your systems, and the longer, the more real work the data captures. Peak headcount and years of history are the first two numbers we look at.

  2. Conversations

    Email, chat, support tickets and call recordings show how decisions actually get made. They are usually the most valuable part of a company's data.

  3. Records that connect

    A job you can follow from first inquiry to quote, work order, invoice and follow-up is worth far more than the same files in isolation.

  4. Specialist work

    Niche industries and technical files — CAD / Revit drawings, source code, engineering specs, expert decisions — are hard to find anywhere else.

  5. Clean ownership

    Data you own outright and can export. Client data, personal information or data held by a partner can often still work — just flag it early.

READY WHEN YOU ARE

Know your fit before you share a thing.

Confirm all four readiness signals above, then complete the high-level qualification intake.

Start the intake