AI Policy

1. Purpose

This AI Policy sets rules for responsible use of the Service's AI capabilities, including ASR, dialogue management, TTS, summarisation, and automated actions.

2. Mandatory transparency to callers

Callers must be clearly informed that they are interacting with an AI system rather than a human, at the latest at the start of the call (Article 50(1) of Regulation (EU) 2024/1689, applicable since 02/08/2026, and Greek Law 5321/2026).

By default, the Service announces this disclosure at the opening of every call. The Customer may adapt the wording in call openings and pre-answer messages, but may not remove it or create any misleading presentation.

Where practically feasible, the Customer provides a clear way to reach a human (e.g., call transfer or callback).

3. Recording disclosures are separate

Transparency that an interaction is AI-driven is distinct from whether a call is recorded.

If the Customer enables recording, the Customer must ensure that callers are informed in advance that the call is recorded and for what purpose, and that the relevant privacy information is provided (Article 4(3) of Greek Law 3471/2006, GDPR). The Customer must not represent calls as recorded if they are not, and must not record without the required legal basis and notices/consents.

4. Quality limitations

AI outputs may be inaccurate, incomplete, or inappropriate. The Company does not guarantee that:

• transcripts will be complete,
• summaries will capture all critical information,
• routing and classification will be error-free.

The Customer must implement reasonableness checks and human review where needed, especially where deadlines, rights, health, finances, or legal consequences are involved.

5. Prohibited AI practices in the Service

The Customer must not use the Service for:

• impersonation, identity deception, or misleading interactions,
• collection of data beyond what is necessary for the Customer's stated purpose,
• fully automated decisions producing legal or similarly significant effects without appropriate safeguards, where applicable,
• emotion recognition or biometric identification/categorisation of callers based on their voice,
• outbound automated promotional calls without the recipient's prior express consent (Article 11(1) of Greek Law 3471/2006),
• training or fine-tuning the Customer's models on third-party data without a lawful basis.

6. Use of Customer Data for training

The Company does not use Call Content or other Customer Data to train general-purpose AI models unless the Customer provides explicit written opt-in consent under agreed safeguards (e.g., anonymisation/pseudonymisation and strict purpose limitation).

7. Human oversight and controls

Where available, the Service provides settings for enabling/disabling recording, retention period, data export, and user access management. The Customer must make appropriate use of these settings.

8. Reports and incident handling

Suspected misleading use, malicious calls, or abuse should be reported to [TRUST & SAFETY EMAIL]. The Company may investigate and take measures, including suspending the account.