Skip to main content

AI Employee Configuration

Table of Contents

  1. Accessing Configuration Settings
  2. Personality Settings
  3. Language and Voice Configuration
  4. Behavior Settings
  5. Call Routing Configuration
  6. Integration Settings
  7. Voicemail Detection
  8. Testing Your Configuration
  9. Configuration Best Practices
  10. Advanced Configuration Tips

Accessing Configuration Settings

To configure an AI employee, navigate to their individual detail page and access the Configuration tab.
  1. Go to the AI Employees section from the main navigation
  2. Click on the AI employee name you want to configure
  3. Select the Config tab from the employee detail navigation

Configuration Structure

The configuration page is organized into logical groups: Core Settings: Fundamental configurations that define the AI employee’s identity and behavior:
  • Personality (greeting and conversation prompt)
  • Languages and Voices
  • Behavior parameters
Call Routing: How calls are handled and transferred:
  • Transfer to Number
  • Transfer to Employee
Integrations: Connections to external systems:
  • Phone Number assignment
  • Calendar integration
  • Voicemail Detection

Personality Settings

Personality is the heart of your AI employee’s identity. It determines how they introduce themselves and conduct conversations.

Greeting Message

The greeting message is what your AI employee says at the very beginning of each call. It sets the tone for the entire interaction. Components of an effective greeting:
  • Introduction of the AI employee or company
  • Clear statement of purpose
  • Invitation to engage
Example greetings: For customer support: “Hello, thank you for calling Acme Corporation. My name is Sarah, your virtual support assistant. How may I help you today?” For appointment scheduling: “Good afternoon! You have reached the appointment desk at City Dental. I am Alex, and I am here to help you schedule your visit. Do you have a specific date in mind?” For sales qualification: “Hi there! This is Jordan from TechSolutions. I am reaching out because you recently expressed interest in our products. Do you have a few minutes to discuss how we might be able to help?”

Greeting Translations

If your AI employee supports multiple languages, you can provide translated versions of the greeting for each language. This ensures callers receive a native-language experience from the first moment. Setting up translations:
  1. Open the Personality settings
  2. Locate the greeting translations section
  3. Enter the greeting text for each additional language
  4. Save your changes

System Prompt

The system prompt defines your AI employee’s overall behavior, knowledge, and conversation style. This is the most powerful and flexible configuration option. What to include in a system prompt: Identity: Who the AI employee represents and their role. “You are a friendly customer service representative for Acme Corporation, a leading provider of home improvement products.” Knowledge base: What the AI employee knows and can discuss. “You can answer questions about our product catalog, store hours, return policies, and delivery options.” Conversation style: How the AI employee should communicate. “Maintain a professional yet warm tone. Keep responses concise but helpful. Ask clarifying questions when the caller’s request is unclear.” Actions: What the AI employee can and should do. “If a customer has a billing issue, offer to transfer them to the billing department. If they want to make a purchase, direct them to our website or nearest store location.” Boundaries: What the AI employee should not do. “Do not make promises about pricing or promotions without verification. Do not provide personal opinions on competitor products.”

AI-Assisted Prompt Generation

The platform can help generate or improve your system prompt:
  1. Provide your company context and the employee’s purpose
  2. The AI assistant suggests prompt language
  3. Review and customize the generated content
  4. Save the final version

Language and Voice Configuration

Configure which languages your AI employee speaks and how they sound.

Primary Language

The primary language is the default for all conversations. It determines:
  • The language of the greeting message
  • The AI model’s default processing language
  • The default voice selection
Selecting a primary language:
  1. Open Languages & Voices settings
  2. Choose from the available language options
  3. This becomes the baseline for all other language configurations

Additional Languages

Expand your AI employee’s reach by adding support for multiple languages. How multilingual support works:
  • The AI employee can detect which language a caller uses
  • They automatically switch to the caller’s language when supported
  • Each language can have its own voice configuration
  • Greeting translations ensure a consistent welcome in any language
Adding additional languages:
  1. Open the Languages & Voices dialog
  2. Click to add languages from the available options
  3. Configure voice settings for each added language
  4. Provide greeting translations

Voice Selection

Each language requires a voice configuration. Voice options include: Voice ID: Select from available synthetic voices that match your brand. Voice Label: A descriptive name for the voice (for your reference).

Voice Parameters

Fine-tune the voice quality with adjustable parameters: Stability: Controls how consistent the voice sounds. Higher values produce more uniform output; lower values allow more expressive variation. Speed: Adjusts how fast the AI employee speaks. Range typically spans from slower (more deliberate) to faster (more energetic). Similarity: Controls how closely the output matches the original voice model. Higher values produce more accurate reproduction. Recommended approach: Start with default values and adjust based on feedback from test calls.

Behavior Settings

Behavior settings control technical aspects of how your AI employee operates during calls.

AI Model Selection

Choose which AI model powers your AI employee’s conversations. Different models offer trade-offs between:
  • Response quality and sophistication
  • Processing speed
  • Cost efficiency
Select the model that best fits your use case:
  • For complex conversations requiring nuanced understanding, choose more capable models
  • For simple, high-volume interactions, more efficient models may suffice

Maximum Call Duration

Set a limit on how long individual calls can last. Why set a maximum duration:
  • Prevents runaway calls from consuming excessive resources
  • Ensures calls do not extend beyond reasonable lengths
  • Provides a safety boundary for edge cases
Choosing a duration:
  • Consider your typical call length needs
  • Add buffer for longer conversations
  • Default is usually 15 minutes (900 seconds)
  • Adjust based on your specific use case

First Message Interruption

Control whether callers can interrupt the AI employee during the initial greeting. Disabled interruption: The greeting plays completely before the AI listens for caller input. This ensures the full introduction is heard. Enabled interruption: Callers can begin speaking at any time, and the AI responds immediately. This creates a more dynamic but potentially chaotic start. Recommendation: For professional settings, consider disabling first message interruption to ensure callers receive complete context before engaging.

Call Routing Configuration

Configure how calls can be transferred during conversations.

Transfer to Number

This feature allows your AI employee to transfer calls to human agents at specific phone numbers based on defined conditions. Configuration components: Phone Number: The destination number where calls should be transferred. Condition: A description of when the transfer should occur. This could be:
  • Caller request (“If the caller asks to speak with a human”)
  • Specific scenarios (“If the issue involves a billing dispute”)
  • Escalation triggers (“If the caller expresses significant frustration”)
Transfer Type:
  • Cold transfer: The AI employee disconnects when the transfer initiates
  • Conference transfer: The AI employee remains briefly to facilitate handoff
Transfer Message: What the AI employee says before transferring, such as “I am connecting you with a specialist who can help with this. Please hold.” Adding multiple transfer rules: You can configure multiple transfer destinations with different conditions:
  • Billing department for payment issues
  • Technical support for product problems
  • Manager escalation for complaints

Transfer to Employee

Transfer calls to another AI employee instead of a human. This is useful for: Specialized routing: Different AI employees handle different expertise areas. Language handoff: Transfer to an AI employee specialized in the caller’s language. Overflow handling: Route to backup AI employees during high volume. Configuration:
  1. Select the destination AI employee from your roster
  2. Define the condition triggering the transfer
  3. Optionally set a transfer message
  4. Configure timing parameters like delay before transfer
Requirements: You need at least two AI employees to use this feature.

Integration Settings

Connect your AI employee to external systems for enhanced functionality.

Phone Number Assignment

Assigning a phone number enables your AI employee to receive inbound calls directly. Current status: Phone number assignment from the configuration page may show as “Coming Soon” with assignment managed through the Phone Numbers settings page instead. To assign a phone number:
  1. Navigate to Settings > Phone Numbers
  2. Purchase or select an available number
  3. Use the assignment dropdown to link it to your AI employee

Calendar Integration

Connect calendars to enable appointment scheduling during calls. Prerequisites:
  • Calendar connections must be established in Settings > Integrations
  • At least one Google Calendar or Microsoft Outlook calendar connected
Configuring calendar access:
  1. Open the Calendar settings in the AI employee configuration
  2. View available calendar connections
  3. Select which calendars this AI employee can access
  4. Configure booking parameters:
    • Duration: Default appointment length
    • Calendar name: Display name for the booking calendar
    • Enable/Disable: Toggle availability for this AI employee
How it works in calls:
  • Caller requests an appointment
  • AI employee checks availability on connected calendars
  • AI employee proposes available times
  • Caller selects a time
  • AI employee creates the calendar event
If no calendars are connected: The configuration will show a note directing you to Settings > Integrations to establish connections first.

Voicemail Detection

Configure how your AI employee handles reaching voicemail during outbound calls.

Why Voicemail Detection Matters

During outbound campaigns, a significant percentage of calls reach voicemail rather than live contacts. Voicemail detection allows your AI employee to:
  • Recognize when voicemail answers
  • Choose whether to leave a message or hang up
  • Deliver a customized voicemail message

Configuration Options

Enable/Disable: Toggle voicemail detection on or off. Behavior when voicemail detected:
  • Leave message: AI employee delivers a pre-configured voicemail message
  • Hang up: AI employee disconnects without leaving a message
Voicemail Message: The message your AI employee leaves on voicemail. Keep it:
  • Brief and clear
  • Including callback information
  • Aligned with the purpose of the call
Example voicemail message: “Hi, this is Alex from TechSolutions. I was calling to discuss your recent inquiry about our services. Please call us back at 555-123-4567 at your convenience. Thank you!”

Voicemail Detection Accuracy

Voicemail detection uses audio analysis to identify voicemail greetings versus live answers. While highly accurate, occasional misdetections can occur:
  • Live callers with formal answering styles may be detected as voicemail
  • Very brief or unusual voicemail greetings may be missed
Monitor your call outcomes and adjust strategies if you notice patterns.

Testing Your Configuration

Before deploying configuration changes, test thoroughly.

Test Call Functionality

Use the test call feature to verify your AI employee’s configuration:
  1. Access the AI employee detail page
  2. Look for test call options
  3. Initiate a call to your own phone
  4. Experience the conversation as a caller would
  5. Note any issues with greeting, responses, or behavior

What to Test

Greeting: Does it sound natural? Is the volume appropriate? Is the message clear? Language: If multilingual, test each language to verify voice quality and greeting translations. Conversation flow: Ask typical questions your callers would ask. Does the AI employee respond appropriately? Transfers: If configured, trigger transfer conditions to verify routing works. Calendar booking: If enabled, attempt to schedule an appointment. Edge cases: Try unexpected questions to see how the AI employee handles them.

Iterative Refinement

Testing is an iterative process:
  1. Make your initial configuration
  2. Test
  3. Identify issues
  4. Adjust configuration
  5. Test again
  6. Repeat until satisfied

Configuration Best Practices

Start with Clear Goals

Before configuring, document:
  • What is this AI employee’s primary purpose?
  • What should a successful call look like?
  • What questions will callers commonly ask?
  • What actions should the AI employee be able to take?

Write Natural Prompts

The system prompt should read naturally:
  • Avoid overly technical language
  • Write as you would instruct a human employee
  • Include context about your business
  • Be specific about desired behaviors

Less Can Be More

Start with simpler configurations:
  • Focused prompt on primary use case
  • Limited transfer rules initially
  • Core language support first
Add complexity only as needed based on real-world usage patterns.

Document Your Configuration

Keep records of:
  • What each AI employee is configured to do
  • Why certain decisions were made
  • Changes made over time
  • Results from different configurations

Monitor and Adjust

Configuration is not “set and forget”:
  • Review call recordings periodically
  • Track success metrics
  • Gather feedback from transferred human agents
  • Adjust based on real performance data

Advanced Configuration Tips

Handling Multiple Scenarios

For AI employees handling diverse caller needs, structure your prompt with clear scenario handling: “When the caller asks about [Topic A], respond with [Approach A]. When the caller asks about [Topic B], respond with [Approach B]. For any other questions, [Default approach].”

Using Context Effectively

Provide relevant business context in your prompt:
  • Current promotions or special offers
  • Recent changes that callers might ask about
  • Seasonal information that affects responses

Balancing Scripted and Flexible Responses

Some responses benefit from consistency (company name, phone numbers), while others need flexibility (addressing unique questions). Your prompt can guide both: “Always mention our website address as www.example.com when directing callers to online resources. For questions about specific product features, provide helpful information based on the caller’s needs.”

Regular Updates

Keep your AI employee’s knowledge current:
  • Update prompts when business information changes
  • Refresh greeting messages seasonally
  • Adjust transfer rules as your team structure evolves

Summary

Effective AI employee configuration requires thoughtful planning and iterative refinement. By understanding each configuration component and following best practices, you can create AI employees that deliver exceptional caller experiences while achieving your business objectives. Remember that configuration is an ongoing process. Monitor performance, gather insights, and continuously optimize your AI employees to meet evolving needs.