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

# Conversational agents: Configuration

> Configuring basic agent settings, prompts, and behavior

Agent configuration defines your agent's identity, behavior, and capabilities. The configuration is split into basic settings that every agent needs and advanced options for customization.

## Basic

For the internal tracking, name your agent in a way that tells you at a glance about its purpose and responsibilities. Name and description are only used for internal organization and navigation.
For example:

* **Name**: Trucking Broker Agent
* **Description**: Replies to incoming broker calls, helps with operational stuff, information requests, etc

<Note>
  Name and description are only used internally, customers never see these
  values as they interact with the agent through its prompt and personality
  settings.
</Note>

### Prompt

The prompt is how you give instructions to your agent. This is where you define what the agent should do, how it should behave, and what knowledge it should draw upon.

**Example basic prompt**:

```
You are a helpful customer service agent for $company_name.
Help customers with their account questions and order status.
If you cannot resolve an issue, use @transfer_to_human to escalate.
```

### Template variables

The prompt editor supports dynamic template variables that get replaced with actual values during conversations:

Built-in variables:

* `$representative_name` - The agent's representative name
* `$company_name` - Your company name
* `$current_date` - Current date and time

**Custom variables**: You can define your own variables in the **Advanced** tab for agent-specific customization.

### Tool references

Reference tools directly in your prompt using the `@` symbol, some built-in tools include:

* `Send DTMF`
* `Send SMS`
* `Send email`

<Tip>
  The prompt editor provides autocomplete when you type `$` for variables or `@`
  for tools, making it easy to discover available options.
</Tip>

## Advanced

### Personalities

Configure how your agent presents itself to customers:

* **Representative name**: The name used to greet the customers, e.g. "Hi, this is Sarah from Acme Corp"
* **Voice**: The voice the agent will use for this personality.

<Note>
  We can support custom voices from **ElevenLabs** and **Cartesia**. Get in touch if you’d like to bring your own.
</Note>

### Language

Configure which languages are enabled for your agent:

* English
* Spanish

When multiple languages are enabled, the agent starts with the set default language, and upon requests switches to the requested language. When the language switches mid-conversation, the underlying voice models automatically change to that language's supported models as well.

### Audio

You can fine-tune the voice experience with the following options:

* **Ambient noise**: Adds light background sounds (e.g., office chatter) to make phone conversations feel more natural and human.
* **Busy sound**: Plays subtle typing or processing sounds to indicate the agent is working or retrieving information.

## Read next

* [Dynamic configuration](/docs/conversational-agents/dynamic-configuration): learn how to configure your agent dynamically at the start of a conversation
* [Context](/docs/conversational-agents/context): learn how you can customize the context provided to a conversational agent