Our approach

Ethically restrained AI. A principle we build by.

Three phrases set the limits on how we build. Kindful is restrained by design: what it refuses to do matters as much as what it offers.

Human wisdom

The awareness is yours

Kindful helps create the conditions in which you can find again what is already there. Human wisdom stays primary.

Machine intelligence

A machine, and it says so

Kindful does not present itself as a person, a therapist or a friend, and it is designed to discourage reliance on itself.

Where you are

In everyday moments

It meets you in the moments of daily life, then points you back to your own experience and relationships.

Safeguards

How we keep people safe.

When Kindful hands over to a person

Crisis contacts are visible throughout the app. When a conversation moves outside Kindful's scope, it says so and points to human support. In our pilots, a named clinician is available for escalation, and adverse experiences are logged and reviewed every week.

Warmth without flattery

AI systems tend to agree with whoever is talking to them. Kindful is designed to stay warm without simply agreeing. It doesn't treat your view of other people as fact, and it leaves room for other perspectives.

Human oversight

Kindful's practices and conversational guidance are designed and reviewed by experienced clinicians and mindfulness teachers.

Your data

Data is handled in line with South Africa's POPIA and with the standards of HIPAA, GDPR and strict medical confidentiality. [Draft wording, to confirm.] In research and pilots it is held under a study code, not a name. Employers never see individual data, and results for groups smaller than five are not reported.