Last updated: November 4, 2026
This policy explains how our concierges use AI and what clients may and may not use them for. It is part of our terms of service.
How our concierges work
A concierge reads the client’s catalog, policies and rules, and uses a large language model to understand questions and write answers in the client’s voice. It is tested on real questions before launch.
It can still make mistakes, so it is built to say when it doesn’t know, to quote prices and policies from the client’s data, and to hand the conversation to a person.
What every concierge does
- Says it is an AI assistant at the start of each conversation, and again whenever someone asks.
- Offers a way to reach a person.
- Answers only about the client’s business, products and services.
- Never asks for passwords, full card numbers, bank details or government ID numbers.
- Never gives medical, legal or financial advice, and never makes decisions about credit, insurance, housing or employment.
What clients may not do
- Sell or promote anything illegal, or age-restricted products such as alcohol, tobacco or cannabis without the age checks the law requires and our written agreement.
- Mislead customers with fake reviews, false urgency, invented discounts, or by presenting the concierge as a person.
- Discriminate, or steer customers, based on race, color, religion, sex, sexual orientation, gender identity, national origin, disability, family status, age or any other protected trait.
- Collect sensitive data, or data from children, through the concierge.
- Send spam or unsolicited messages, or use WhatsApp against Meta’s Business Messaging and Commerce policies.
- Use the concierge to harass, threaten or deceive anyone, or to produce sexual, violent or hateful content.
- Try to get around the concierge’s safety rules, reverse engineer our software or use it to build a competing product.
- Run it as a general-purpose chatbot unrelated to the client’s business.
Rules by industry
Real estate
Concierges follow fair housing rules in the United States and anti-discrimination rules in Brazil. They describe properties, not the people who live nearby, and never steer anyone.
Clinics, med spas and health
No diagnosis, no judgment on whether a treatment suits someone, and no promised results. Medical questions go to a licensed professional. Clients follow the advertising rules of their professional boards.
Restaurants and hotels
Allergy questions always go to the kitchen. Rates, deposits and cancellation terms are quoted from the client’s own policy.
Car dealers
Payment figures are labeled estimates that show the terms used, never a promise of credit. Credit decisions stay with the lender and the dealer.
Watches, jewelry and luxury
Condition, authenticity and papers are confirmed by a specialist, not by the concierge.
Laws on AI disclosure
Several US states, including California, Utah, Maine and New Jersey, require businesses to tell customers when they are dealing with a bot, and Brazilian consumer law requires clear information. Our concierges disclose by default, and clients may not turn that off.
Reporting a problem
Report a wrong answer, misuse or a security issue to hello@createyourai.co. We may pause a concierge that breaks this policy or puts people at risk, and we tell the client why.