1. Transparency
Article 50 AI Act. From the first message, the user is clearly informed that they are interacting with an artificial intelligence system, not a person. This notice is visible and unambiguous.
2. Human oversight
The AI receptionist assists the salon but does not replace it. The salon keeps control: it can take over at any time, edit or cancel a response, and validate appointments. No decision with legal effect is taken without possible human intervention.
3. Limits and errors (hallucinations)
Like any generative AI system, the assistant may produce inaccurate or incomplete responses (“hallucinations”). Critical information (hours, prices, availability) must be verified by the salon. Bellisy continuously improves the quality and reliability of the system.
4. Prohibited uses
- Use AI for discriminatory or deceptive decisions.
- Pass the AI off as a human.
- Process sensitive data for profiling.
- Generate illegal, hateful or misleading content.
5. Training data and providers
Bellisy relies on selected AI model providers. Client and salon data is not used to train third-party models without an appropriate legal basis. Sensitive data is excluded from any training.
6. Risk management
Bellisy regularly assesses AI-related risks (bias, reliability, security, privacy) and implements mitigation measures: content filters, guardrails, interaction logging and feedback mechanisms.
7. Complaints
Anyone can report an AI-related issue (inappropriate response, error) to support@bellisy.fr. Reports are reviewed and used to improve the system.