AI Use Policy

This policy covers the "AI assistant", an optional service on the Enterprise plan and during the trial that a business turns on itself. The assistant answers the business's customers in WhatsApp on the business's number, on the business's chat page on wgtk.net, and in the test window inside the dashboard.

No other part of the platform uses AI to make decisions about people.

• Answers customers' questions about the business from what the business entered: services, prices, durations, hours, location, FAQs and policies.
• Books an appointment or a table at a time that is really free. The platform's booking engine checks the time and the staff member before confirming any booking, so the assistant cannot book a taken slot.
• Hands the customer to a person at the business when asked or when it cannot answer.

• It never claims to be human. If asked, it says it is the business's AI assistant and offers to pass the customer to a person.
• It gives no medical, legal or financial advice, makes no diagnosis and predicts no case outcome. The business's specialist answers those in the appointment.
• It invents no price, discount or offer that is not in the business's data.
• It never takes card details. Payment happens on a secure page at the payment provider.
• It makes no decision with a legal or financial effect on the customer, and does not profile or rate customers.

We tell customers that a reply may be written by an AI assistant, in the privacy policy and in the conversation itself when asked. A customer can ask for a person at any time, and the business can stop the assistant in any conversation and reply itself.

• The business's information as entered in the dashboard, and the free times from its schedule.
• The customer's messages in the current conversation, and their name and phone number, to do what they ask.

Conversations and customer data are not used to train AI models. The retention periods and data subject rights in the privacy policy apply to them.

The assistant runs on a language model from Anthropic (the Claude family), which we call directly under our own agreement with Anthropic. Processing happens outside Saudi Arabia, with only the minimum data needed for the reply, encrypted, under contract terms that forbid training on the data, and within the transfer rules of the Personal Data Protection Law.

• The business reviews its conversations in the inbox, can switch any conversation from the assistant to itself, or turn the assistant off entirely from its page in the dashboard at any time, so every message reaches the inbox for a person to answer.
• We can stop the assistant at once, for one business or for all of them, if it does not work as it should or puts customers at risk; messages then reach each business inbox.
• When a customer asks for a person, the business's team is alerted at once with the customer's details and request.
• WGTK support reads conversations only when needed: when the business reports a conversation, or to fix a fault in the service. Access is time-limited, phone numbers are masked, and every access is logged.

We classify the assistant as a low-risk use, because it provides customer service and appointment booking and makes no decision with a significant effect on people. We monitor it continuously: failed replies, conversations waiting for a person, the bookings it made, and technical errors, and we reassess the risk whenever the way it works changes materially.

Wgtk Orbit Company is responsible for running the assistant and for this policy. For any remark or complaint about a reply, or a request about your data, email us at the address below; we answer within 30 days.

We may update this policy when the way the assistant works changes, and publish the updated version here with its date.