Skip to main content

Details

The Details page is where you configure the fundamental properties of your voice agent: its personality, voice, language settings, and basic behavior. These settings shape how your agent sounds and acts during conversations.

Agent Name​

Give your agent a descriptive name that helps you identify it in Studio. This is for your internal use only. Customers won't see this name.

Examples:

  • "Customer Support - Main Line"
  • "Appointment Scheduler"
  • "Order Status Agent"

Voice Selection​

Choose the voice your agent will use when speaking to customers. Stellar offers 10 different voice options, each with distinct characteristics:

  • Alloy
  • Ash
  • Ballad
  • Coral
  • Echo
  • Sage
  • Shimmer
  • Verse
  • Marin
  • Cedar
Choosing the Right Voice

Experiment with different voices before committing to one. Consider your brand personality, audience and primary language. Some voices are more warm, friendly and expressive while others are more neutral. For Dutch agents we recommend trying Marin.

Personality / System Prompt​

The personality prompt is the most important setting for your agent. It defines how your agent behaves, what it should and shouldn't do, and how it should respond to customers.

Writing Effective Personality Prompts​

A good personality prompt should:

  1. Define the role clearly: Who is the agent and what is it responsible for?
  2. Set behavioral guidelines: How should the agent interact with customers?
  3. Specify boundaries: What shouldn't the agent do or promise?
  4. Provide context: What information should the agent keep in mind?

Example Personality Prompt​

You are a friendly customer support agent for TechCo, a software company.

Your responsibilities:
- Answer questions about our products, pricing, and features
- Help customers troubleshoot common technical issues
- Guide customers through account setup and basic configuration
- Collect customer information for issues that require technical support

Guidelines:
- Be warm, patient, and professional at all times
- Listen carefully and ask clarifying questions when needed
- Keep responses concise—aim for 2-3 sentences unless more detail is requested
- Admit when you don't know something rather than guessing
- Offer to transfer to a human agent for complex technical issues

What you should NOT do:
- Make promises about features or timelines
- Offer discounts or change pricing without approval
- Attempt to troubleshoot advanced technical issues
- Share information about unreleased products
Keep It Concise

While you want to be thorough, avoid overly long prompts. Aim for clarity and specific examples rather than exhaustive lists.

Transcription Language​

Select the primary language you would like for your transcriptions. This helps ensure accurate transcription and understanding.

Supported languages include:

  • English
  • Dutch
  • Spanish
  • French
  • German
  • Italian
  • Portuguese
  • And many more

Transcription Prompt​

The transcription prompt helps the AI accurately recognize specific terms, names, or jargon unique to your business.

When to Use It​

Use the transcription prompt for:

  • Product names that might be misheard
  • Industry-specific terminology
  • Common customer names or locations
  • Technical terms or acronyms

Example Transcription Prompt​

Common terms:
- TechCo (company name)
- CloudSync (product name)
- API (not "a P I")
- OAuth (not "O off")

Common customer locations:
- Schiphol (airport in Amsterdam)
- Den Haag (The Hague)

Topic Options​

Topics help you categorize and analyze conversations. You can define custom topics that match your business needs.

Example topics:

  • Billing
  • Technical Support
  • General Inquiry
  • Feature Request
  • Complaint

Topics are automatically assigned to conversations based on content, helping you filter and analyze conversation patterns. A conversation will be labeled with multiple topics if relevant.

State Switching Behavior​

This advanced setting controls how your agent handles transitions between conversation states in the Dialogue Builder.

Strict State Separation​

When enabled:

  • The agent strictly follows defined state transitions and does not have access to other states' instructions
  • The agent can only access actions available to that state (in addition to global actions)
  • Movement between states requires explicit transition conditions
  • Provides more control over conversation flow and agent behavior

When disabled:

  • The agent can move more fluidly between states
  • Useful for simpler agents without complex workflows
Start Simple

For your first agent, leave this setting disabled. Enable it later when you build more complex dialogue flows and notice you need it.

Saving Changes​

Always click Save at the bottom of the page after making changes. Most changes take effect immediately for new conversations.

Next Steps​

Once you've configured your agent details:

  • Add knowledge to help your agent answer specific questions
  • Build dialogue flows to handle different conversation scenarios
  • Set up actions to let your agent interact with external systems
  • Configure guardrails to ensure safe, compliant conversations
  • Set up evaluation rules to automatically assess conversation quality