Skip to content
KetDataValue your records

For AI developers

Licensed, de-identified records of real company work, with a clean rights chain and one counterparty.

What you get

Real working records from established companies: email, chat, tickets, code and documents, linked across years.

See the dataset types

How it's prepared

  • Our specialist de-identification partner extracts the records once and de-identifies them.
  • You receive the de-identified master copy and a de-identification certificate.
  • KetData never holds raw data.

Licence terms in brief

The licence

  • Non-exclusive by default; exclusive at a premium.
  • Perpetual.
  • To train, evaluate and develop AI models. Derived models are allowed.

Not allowed

  • Public release.
  • Resale or sublicensing of raw records.
  • Any attempt to re-identify individuals.
  • Decisions about named individuals.

How it works

The company is named only after you have signed a non-disclosure agreement.

  1. Non-disclosure agreement

    You sign before we name any company.

  2. Profile without the company's name

    Sector, size, history, systems and volumes.

  3. Sample review

    You review a sample before you commit.

  4. Licence

    One licence, with KetData as your counterparty.

  5. Delivery

    The de-identified master copy and its certificate.