This page explains how Azul Interactiv LLC uses artificial intelligence on this website, what happens to what you type into our chat, and the limits of what that chat tells you. It applies to the chat assistant on our contact and booking pages.
1. You are talking to an AI, not a person
The chat assistant on this site is artificial intelligence, not a member of our team. Nobody reads your messages as you type them, and no person reviews the assistant's replies before you see them.
This website uses an AI model, Claude Haiku 4.5, made by Anthropic PBC, to hold that conversation, understand what you are trying to build or fix, and decide whether Azul Interactiv is a fit for the work.
You are never required to use the chat. Emailing [email protected] reaches a person directly and involves no AI processing of your message.
2. What is sent to our AI provider
When you type into the chat, your messages are transmitted to Anthropic via their API for processing. Because it is an open conversation rather than a set of form fields, this includes whatever you choose to tell it about your business, your systems, your budget constraints, and your project.
Please do not enter passwords, API keys, financial account details, or confidential information belonging to a third party. None of it is needed to find out whether we can help you.
See Anthropic's Privacy Policy for how they handle data submitted through their API.
3. What we receive — including if you leave partway through
When the conversation reaches a conclusion, we receive your contact details and a summary of what you told us.
If you give us an email address and then leave the page before finishing, we also receive a copy of the conversation up to that point. We do this so that an enquiry you started does not simply vanish, and so we can follow up. If you would rather that did not happen, do not enter an email address into the chat — email us directly instead.
4. Training
We do not use your conversations to train or improve AI models, and we do not sell them.
Anthropic's Commercial Terms state that inputs and outputs submitted through their API are not used to train their models by default. Because your conversation is not used for training, there is no training opt-out to offer. We link the source rather than paraphrasing it, so you can rely on the provider's own wording: Anthropic Commercial Terms of Service.
5. AI-generated content may be wrong
AI-generated content may contain errors or inaccuracies. The assistant's replies are automated and unreviewed. Anything it says about what we can build, how we work, whether your project is a fit, or what a system would do, may be incomplete or simply incorrect.
Nothing it produces is professional advice, a quote, a proposal, or an offer, and it does not create a contract or a client relationship. Use your own judgment, and confirm anything that matters with Javier on the call, before relying on it for a business, financial, or legal decision.
6. Commercial context
The chat assistant is operated by Azul Interactiv to evaluate prospective clients for our own paid services, and to book qualified enquiries into a call. Its questions and its conclusions serve that purpose.
It does not compare us against other providers and does not give impartial advice about whether you should engage us. If it tells you we are a good fit, read that as what it is: an assessment made by the company selling the work.
7. Questions
If you have questions about how we use AI, or want a copy of what you told the chat assistant, email [email protected].