How-To: Agent

An Agent is the core entity that combines voice, behavior, and context, enabling personalized and smooth AI-powered conversations with customers.

​

Steps to Create an Agent

​

1. Agent Name

Define the name of your Agent.

​

2. Initial Message

Set the initial message your Agent will send when the call is connected. Any {variable_name} included will be replaced with the corresponding context. If there’s no match, the Agent will simply say the variable name.

​

3. Voice

Select the voice options created and available for your Agent.

​

4. Prompt Library

Choose from the pre-made prompts available for your Agent.

​

5. Webhooks

Select webhooks that your Agent can trigger.

​

6. Actions

Select the available actions your Agent can perform.

​

Advanced Settings

• Language
  Set the preferred language for your Agent.

• Model Name
  Choose from the available models:

  • GPT-4
  • GPT-4 Turbo
  • GPT-4 Omni
  • GPT-4 Omni Mini
  • GPT-3.5
  • GPT-3.5 Turbo

Note: Latency may vary depending on the model selected.

• Conversation Speed
  Adjust the waiting time to ensure the Agent allows the human speaker to finish their sentence. A lower value makes the Agent wait longer, while a higher value allows the Agent to speak during the human’s speech.

• Prompt
  If a prompt is selected, the library will be loaded and updated with any changes made.

You can explore our API’s Prompt Creation feature here.

Ready to continue? Check out our Actions Guide to learn more.