Skip to main content

Overview

Agent configuration is the foundation of creating effective voice AI agents. This guide covers everything you need to know about setting up, customizing, and optimizing your agents.

Creating Your First Agent

Step 1: Choose a Starting Point

  • Template: Start with a pre-built template from the Agent Templates Library
  • Blank Agent: Create from scratch for complete customization
  • Clone: Duplicate an existing agent as a starting point

Step 2: Basic Information

  • Agent Name: Give your agent a descriptive name
  • Description: Add a brief description of the agent’s purpose
  • Category: Select the appropriate category (Sales, Support, etc.)

Step 3: Configure Core Settings

  • Voice Selection: Choose a voice from the library or clone your own
  • Language: Select primary language and additional languages
  • Personality: Define the agent’s tone and style

Step 4: Set Conversation Goals

  • Primary Goal: Define the main objective (e.g., “Schedule appointments”)
  • Success Criteria: Specify what constitutes a successful call
  • Key Metrics: Identify metrics to track (conversion rate, duration, etc.)

Step 5: Test and Deploy

  • Playground Testing: Test in the Tabbly Playground
  • Iterate: Refine based on test results
  • Deploy: Activate for live calls

Agent Settings and Parameters

General Settings

Agent Name & Description
  • Internal name for your reference
  • Description helps team members understand the agent’s purpose
Status
  • Active: Agent is live and handling calls
  • Paused: Temporarily disabled
  • Draft: Still in development
Priority
  • Set call routing priority if multiple agents are available
  • Higher priority agents handle calls first

Conversation Settings

Greeting
  • Customize the opening greeting
  • Set tone (formal, friendly, professional)
  • Include dynamic variables (customer name, time of day)
Conversation Flow
  • Define conversation structure
  • Set branching logic for different scenarios
  • Configure question sequences
Response Style
  • Concise: Short, direct responses
  • Detailed: Comprehensive explanations
  • Adaptive: Adjusts based on caller needs

Behavior Settings

Interruption Handling
  • Allow or prevent caller interruptions
  • Set interruption sensitivity
  • Configure how to handle mid-sentence interruptions
Silence Detection
  • Set timeout for silence
  • Configure prompts for silence
  • Handle “dead air” scenarios
Error Handling
  • Define responses for misunderstandings
  • Set retry logic for failed actions
  • Configure fallback behaviors

Configuring Agent Personality

Personality Traits

Formality Level
  • Formal: Professional, business-like tone
  • Casual: Friendly, conversational tone
  • Adaptive: Adjusts based on caller’s style
Enthusiasm
  • High: Energetic and enthusiastic
  • Moderate: Balanced and professional
  • Low: Calm and measured
Empathy
  • High: Shows understanding and compassion
  • Moderate: Professional empathy
  • Low: Factual and direct

Voice Characteristics

Pace
  • Fast: Quick speech for efficiency
  • Normal: Standard conversational pace
  • Slow: Deliberate and clear
Tone
  • Warm: Friendly and welcoming
  • Neutral: Professional and balanced
  • Authoritative: Confident and knowledgeable
Accent & Dialect
  • Select regional accent if applicable
  • Match your target audience’s preferences
  • Consider brand voice alignment

Setting Conversation Goals

Primary Goals

Lead Qualification
  • Identify qualified leads
  • Collect key information
  • Score leads based on criteria
Appointment Scheduling
  • Schedule appointments
  • Confirm availability
  • Send calendar invites
Customer Support
  • Resolve customer issues
  • Answer frequently asked questions
  • Escalate when needed
Sales
  • Present products/services
  • Handle objections
  • Close deals

Success Criteria

Define what makes a call successful:
  • Conversion: Appointment scheduled, sale made, issue resolved
  • Information Collected: Required data gathered
  • Customer Satisfaction: Positive interaction
  • Efficiency: Call completed within target duration

Key Performance Indicators (KPIs)

Track these metrics:
  • Conversion Rate: Percentage of successful calls
  • Average Call Duration: Time per call
  • First Call Resolution: Issues resolved in one call
  • Customer Satisfaction: CSAT scores
  • Transfer Rate: Calls escalated to humans

Defining Success Criteria

Quantitative Metrics

Conversion Metrics
  • Appointment booking rate
  • Lead qualification rate
  • Sales conversion rate
  • Issue resolution rate
Efficiency Metrics
  • Average call duration
  • Calls per hour
  • Cost per conversion
  • Time to resolution
Quality Metrics
  • Customer satisfaction (CSAT)
  • Net Promoter Score (NPS)
  • First call resolution rate
  • Call quality scores

Qualitative Criteria

Conversation Quality
  • Natural flow and engagement
  • Appropriate responses
  • Professional demeanor
  • Problem-solving effectiveness
Customer Experience
  • Positive interaction
  • Clear communication
  • Helpful and informative
  • Respectful and courteous

Agent Performance Optimization

Monitoring Performance

Analytics Dashboard
  • Real-time performance metrics
  • Historical trends and patterns
  • Comparative analysis
  • Custom reports
Call Review
  • Listen to call recordings
  • Review transcripts
  • Identify improvement opportunities
  • Track success patterns

Optimization Strategies

A/B Testing
  • Test different greetings
  • Compare conversation flows
  • Evaluate response variations
  • Measure impact on metrics
Prompt Refinement
  • Refine prompts based on results
  • Add context and examples
  • Improve clarity and specificity
  • Update based on feedback
Continuous Improvement
  • Regular performance reviews
  • Update based on new data
  • Incorporate best practices
  • Stay current with platform updates

Best Practices

Begin with a basic configuration and add complexity gradually. Test each change before adding more.
Clearly define what success looks like for your agent. This guides all configuration decisions.
Use the Playground to test various scenarios before going live. Include edge cases and difficult questions.
Regularly review performance metrics and make adjustments. Optimization is an ongoing process.
Collect feedback from callers and your team. Use insights to improve agent performance.

Next Steps