Our approach

How we build clinical AI responsibly: clinical safety, regulation, data protection, fairness and evidence, from the start.

Our principles

  1. The clinician decides.Every output is a draft or a prompt. Nothing happens without a clinician.
  2. Every suggestion is explained and cited.Each prompt shows its reasoning and links to the guidance behind it.
  3. Abstain rather than guess.When the evidence isn't there, the platform says so and shows what's missing.
  4. Test for fairness.We break results down by age, sex, ethnicity and deprivation.
  5. Protect data by design.UK GDPR, a DPIA before any real data is used, and UK-hosted, approved environments.
  6. Publish what we learn.Including what didn't work.

In more detail