---
title: Step 2 — Define Agent
slug: step-2-define-agent
docTags: 
createdAt: 2026-03-23T21:27:33.561Z
---

Provide the basic configuration details for the agent.

### Agent Name

Enter a name that clearly identifies the purpose of the agent.

**Example**: Healthcare Appointment Assistant

### Description

Provide a short description explaining what the agent does.

**Example**: Helps users identify symptoms and book medical appointments quickly through a simple and efficient interaction.

![](https://api.archbee.com/api/optimize/FEqjvWZnyymr_PEhXLLlG/qjZ3tk0OjwZjPUiQk-VMF_image.png)

### Enable Voice

Enable the **Voice** option to allow the agent to interact using voice.

### Agent Initiates Conversation

Enable this option if the agent should start the conversation automatically when the interaction begins.

![](https://api.archbee.com/api/optimize/FEqjvWZnyymr_PEhXLLlG/K-LSVK-nDlm4I_EQSuGHC_image.png)

### Voice Mode

Select how the voice interaction will be handled.

Available options may include:

**Speech-to-Speech**

The system processes speech input and returns speech output directly.

**Speech-to-Text and Text-to-Speech**

Speech is converted into text, processed by the AI model, and then converted back into speech.

RECOMMENDED: **Speech-to-Text and Text-to-Speech**

### Speech-to-Text Provider

Select the provider responsible for converting user speech into text.

Example providers:

- Whisper
- Azure
- Google
- Sarvam
- Cartesia

RECOMMENDED: **Sarvam**

### LLM Provider

Select the AI model provider that will generate responses.

Example providers:

- Groq
- Simplismart
- Gemini

RECOMMENDED: **Gemini**

### Text-to-Speech Provider

Choose the provider used to convert text responses into voice output.

Example providers:

- Sarvam
- Google
- ElevenLabs
- Cartesia

RECOMMENDED: **Cartesia**

![](https://api.archbee.com/api/optimize/FEqjvWZnyymr_PEhXLLlG/GI5RYpyxcuFZ4wfdvnhF6_image.png)

### Voice Selection

Select a voice for your agent from the available options.

- Choose a voice based on tone and use case (e.g., confident, friendly, professional).
- Preview the available voices (both **male and female**) before selecting.
- Pick a voice that best matches your agent’s purpose and user experience.

After completing these settings, click **Next** to continue.

## 📚 Next Steps

Continue to: [Step 3 — Workflow Setup](docId\:AJoC0HgUV3DHBNchPjq9Y)