AI transparency

1. What StayHelp is

StayHelp is a software service (SaaS) that lets hotels, B&Bs and short-term rentals offer their guests a digital concierge. Guests scan a QR code and chat — on Telegram or on this website — with an artificial intelligence system that answers questions about the property (check-in times, Wi-Fi, breakfast, parking, house rules and similar). This page is the transparency notice required by Article 50 of Regulation (EU) 2024/1689 (“AI Act”) and complements our Privacy Policy.

2. You are always told when you are talking to an AI

Every conversation opens with a clear notice that you are interacting with an artificial intelligence and not with a person. The notice is shown at the latest at the first interaction, in the language of the guest, and property owners cannot remove or hide it: it is appended automatically even when a property sets its own welcome message. In the Telegram chat you can type /ai at any time to see the full notice again.

3. How the system works

Knowledge base. Each property enters the information about its own accommodation (times, services, rules). The assistant is instructed to answer only from that information and from the property address; when it has no reliable answer it says so and passes the question to a human instead of guessing.
Model. Answers are generated by a general-purpose large language model provided by a third party (currently DeepSeek), accessed through its API. The property’s knowledge base and the recent messages of the conversation are sent to the model to produce the answer.
No training on your conversations. We do not use guest conversations to train or fine-tune any model, and we do not sell them.

4. What the system does not do

The assistant does not take any automated decision producing legal or similarly significant effects on you (Art. 22 GDPR). It does not profile guests, does not perform biometric identification or categorisation, does not recognise or infer emotions, does not use subliminal or manipulative techniques, does not exploit vulnerabilities and does not carry out social scoring — none of the practices prohibited by Article 5 of the AI Act. It does not decide bookings, prices, refunds or access to the property.

5. Limits and accuracy

Generative AI can produce answers that are wrong, outdated or incomplete. The information given by the assistant is indicative and does not replace what the property tells you directly. Do not rely on it for emergencies or for medical, legal or financial matters: in an emergency call the local emergency number (112 in the European Union). If an answer looks wrong, ask for a human (see below) and tell the property, so the underlying information can be corrected.

6. Human oversight

A person is always reachable. You can type /human in the chat at any moment and your request is forwarded to the property, which replies to you personally. In addition, when the assistant is not confident about an answer it escalates the question to the property automatically instead of inventing one. The property can read the conversations, correct the knowledge base and reply in place of the assistant at any time.

7. Your data

The property you are staying at is the data controller of the guest conversations; Touch Informatica Srls acts as processor on its behalf. Messages are stored to allow the property to answer you and are deleted after 90 days by default. The AI model provider processes the message content in order to generate the answer and may do so outside the European Union, on the basis of standard contractual clauses. Full details, retention periods and how to exercise your GDPR rights are in the Privacy Policy.

8. Roles under the AI Act

Touch Informatica Srls (Spoltore, Italy — VAT IT02431740683) is the provider of the AI system. The hotel, B&B or rental that activates StayHelp is the deployer (Art. 3 AI Act) and undertakes, in our Terms of Service, to keep the AI notice active, to keep the information given to guests accurate, to handle escalations and to ensure that the staff using the system has an adequate level of AI literacy (Art. 4 AI Act).

9. Risk classification

We assessed StayHelp against the AI Act: it does not implement any prohibited practice (Art. 5); it is not a high-risk system (Art. 6 and Annex III — an information assistant for hospitality is not among the listed use cases) and it is not a safety component of a product. It is therefore an AI system subject to the transparency obligations of Article 50 and to the AI literacy obligation of Article 4. We repeat the assessment whenever we add significant features.

10. Contact and complaints

For any question about the use of AI in StayHelp, or to report an inaccurate or inappropriate answer, write to [email protected]. You may also complain to the national market surveillance authority for the AI Act (in Italy: Agenzia per la Cybersicurezza Nazionale and AgID) and, for personal data matters, to your data protection authority (in Italy: Garante per la protezione dei dati personali).