How the AI decides what to say

When someone writes to you, Usetta does this, in about the time it takes you to read the message:

1. It works out what the person wants. A price question, a request to book, an objection, a request for a person, or small talk.
2. It reads what you taught it. Your AI Agent settings, your knowledge documents, your listings, and what this person has said before.
3. It writes a reply in the tone you chose, in the language the customer used, and ends with one clear question so the conversation keeps moving.
4. It checks the reply against your rules, including anything on your Do not say list, before anything is sent.
5. It sends the reply, or saves it as a draft, depending on the channel’s mode.

It never guesses facts

If the answer is not in what you gave it, the AI says it will check and, where it makes sense, hands the chat to you. Adding the missing answer to the AI Agent page fixes it for every future customer.

It tells people it is an AI

The first reply to a new person says they are talking to your business’s AI assistant.

It knows when to step back

The AI hands the chat to a person when the customer asks for one, or when the message needs a human decision. The chat appears under Escalations and you are notified. See When the AI hands a chat to you.