> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aide.app/llms.txt
> Use this file to discover all available pages before exploring further.

# How Aide tells customers it is an AI

> The Aide AI Agent never claims to be a person.

The Aide AI Agent never claims to be a person. It says it is an AI when a customer asks, Aide Chat labels it as an AI agent, and a reply that claims to be human is not sent. None of this can be switched off or reworded.

## When a customer asks "am I talking to a person?"

Aide answers with a fixed line, then replies to the rest of the message:

"I'm \{agent name}, an AI agent for \{your brand}. I can bring in a person from our team any time you like."

* In chat, the line comes first in the reply. By email, it goes under the greeting.
* If the customer asked nothing else, Aide follows the line with a short question about how it can help.
* A customer who then asks for a person is handed to your team (see [How Aide hands a conversation to your team](https://docs.aide.app/aide-ai-agent/handoff.md)).

### Where the name comes from

The agent's name is the **Name** under **AI Agent > Tone & identity**. If the name is empty, the line says "Aide". The brand in this line is your workspace's name.

### Languages

The line is sent in the language the customer writes in. The fixed wording exists in English, French and German. In any other language, Aide answers the question in its own words, in one short sentence: that it is an AI agent, and that a person from your team can take over whenever the customer likes.

## What doesn't change it

* **Instructions and procedures.** Neither can change the line or remove it.
* **Refers to itself as**, under **Tone & identity**. The fixed line says "AI agent" whatever is picked there.
* **AI disclosure**, under **Style** in **Tone & identity**. It is part of the agent's style text. The fixed line is sent whether it is on or off.
* **Your own widget text.** The lines that say the agent is an AI can't be edited (see [Change the chat widget's text and languages](https://docs.aide.app/aide-chat/widget-languages.md)).

## The AI labels in Aide Chat

Aide Chat shows that the agent is an AI before the customer has to ask:

* **In the header**: the agent's name followed by "AI agent".
* **Before the first message**: "\{agent name} is an AI agent for \{your brand}. This conversation is stored so we can help you and improve support. You can ask for a person at any time."
* **On proactive messages**: "AI agent", or "AI agent for \{your brand}".

The brand in these labels is the **Brand name** under **Aide Chat > General** (see [Customise how Aide Chat looks](https://docs.aide.app/aide-chat/widget-appearance.md)). They follow the language the widget is shown in.

## A reply that claims to be human is not sent

Before a reply goes out, Aide checks it for a claim to be a person, such as "I'm a real person" or "I'm not a bot", in English, French, Spanish and German. A reply that makes one is rejected and written again without the claim.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.