Data processing — impact assessment

We use two solutions that deserve special attention: quiet behavioural profiling to tailor communication, and an AI voice assistant. Below we explain what we do and what we deliberately do not do.

We analyse behaviour signals (e.g. which sections you read, where you came from) to tailor the tone of the first conversation. The result stays with us only.

We do not recognise emotions, use biometrics, perform social scoring or make automated decisions affecting price. A profile is a hint, not a verdict. You can opt out (see the privacy policy).

When launched, the assistant’s opening line discloses that you are talking to AI and offers to hand the call to a human.

We do not store biometric voice data of the caller. We may analyse the transcript for quality purposes.

The cloned voice used by the assistant belongs to a member of our team, used with their consent. We do not clone third-party voices without written consent.

Our solutions fall under the limited (transparency) risk category. We label AI-generated content, disclose use of the voice assistant and list the models used in the privacy policy.

Questions about data processing: kontakt@gecko-labs.pl.