Working with frontier AI teams and data-rich companies
Start a conversation

Bring us a capability gap, or the data that can close it.

Tell us whether you are building frontier models, evaluating a system, or exploring the value of proprietary company data. The first conversation can stay high-level.

Submissions are saved in this browser only and are not transmitted anywhere. Review or export them on the submissions page. Connect the form to your CRM or backend endpoint before launch.

Thank you.
Your enquiry has been saved in this browser. View saved submissions ↗
For frontier AI teams

Describe the model failure.

The most useful first note includes the target task, the current baseline, the tools involved and what a successful run must leave correct.

  • Target capability and failure examples
  • Model or agent configuration
  • Training, evaluation or both
  • Runtime and security constraints
Review the lab offering
For data owners

Describe the operating history.

A schema, inventory or controlled sample is enough for the first assessment. Do not email raw data before a secure process is agreed.

  • Industry and workflow
  • Systems and years represented
  • Corrections, approvals and outcomes
  • Ownership and known restrictions
Review data partnerships