Skip to main content

​
Overview

Your agent’s voice is a critical part of the customer experience. The right voice can build trust, convey professionalism, and align with your brand identity. itellicoAI streams live catalogs from ElevenLabs, Microsoft Azure Neural voices, and Cartesia so you can choose high-quality audio without manual uploads.
Voice selection happens under the Voice tab in your agent configuration. Changes apply immediately.

​
Voice Providers

ElevenLabs

Premium AI voices with exceptional naturalness and emotional range.Why it works:
  • Ultra-realistic, nearly indistinguishable from human speech
  • Strong emotional range for customer service
  • Consistent quality across all content
  • Low latency for real-time conversations
Best for:
  • Customer-facing agents where voice quality is critical
  • Brand-sensitive applications
  • Use cases requiring emotional intelligence
Popular voices:
  • Rachel: Warm, professional American female
  • Adam: Confident, clear American male
  • Susi: Natural, professional German female (recommended for German agents)
  • Antoni: Calm, reassuring male
ElevenLabs voices support advanced settings like stability and similarity boost—configure in Voice Settings.
Enterprise-grade voices with massive language coverage.Why it works:
  • 100+ languages and locales
  • EU hosting available for GDPR compliance
  • Consistent, professional quality
  • Predictable enterprise pricing
Best for:
  • Multilingual agents (one provider for all languages)
  • Enterprise compliance requirements
  • High-volume applications with cost constraints
  • Global deployments
Popular voices:
  • en-US-JennyNeural: Natural American female
  • en-GB-SoniaNeural: British female, professional
  • de-DE-KatjaNeural: German female, authoritative
Voice tiers:
  • Standard Neural: High-quality, cost-effective
  • Neural HD: Enhanced quality
  • Custom Neural: Train your own voice (enterprise only)
Trade-offs:
  • Slightly less emotional nuance than ElevenLabs
  • Best for factual, professional conversations
Ultra-low latency voices optimized for conversational AI.Why it works:
  • Optimized for sub-second turn taking
  • Expressive, energetic deliveries
  • Modern sound tuned for interactive agents
Best for:
  • Speed-critical web experiences
  • A/B testing alongside ElevenLabs
  • Latency-sensitive applications
Trade-offs:
  • Smaller catalog (primarily English)
  • Fewer customization options
Need another TTS provider (Google Cloud, Amazon Polly)? Contact your success manager—we’ll add it to the catalog.

​
Choosing the Right Voice

​
Selection Framework

Choose based on your requirements:Quality-first? → ElevenLabs (most natural, emotional range)Need specific language? → Azure Speech (strong language coverage, 100+ languages)Speed-critical? → Cartesia (ultra-low latency)EU compliance? → Azure (EU-hosted options)
Industry context:
  • Healthcare: Empathetic, professional, reassuring
  • Sales: Confident, enthusiastic, persuasive
  • Technical Support: Patient, clear, knowledgeable
  • Hospitality: Warm, welcoming, friendly
Accent considerations:
  • Local accents build rapport with local customers
  • Neutral accents work for global audiences
  • Filter by region/locale in voice library
Testing process:
  1. Preview ElevenLabs voices using the play button
  2. Shortlist 3-5 voices that match your criteria
  3. Deploy each to a test agent
  4. Call and test with realistic scenarios
  5. Have team members evaluate
Evaluation criteria:
  • Brand fit and personality match
  • Clarity and naturalness
  • Performance with industry terminology
  • Pleasant to listen to in 5+ minute conversations

​
Voice Library Features

The voice library provides search and filtering to find the right voice quickly: Search by:
  • Voice name (e.g., “Sarah”, “Professional Male”)
  • Provider (ElevenLabs, Azure, Cartesia)
  • Gender (male, female, neutral)
  • Language or locale code (en-US, es-ES, de-DE)
  • Accent or region (British, Australian, American)
Filter by:
  • Provider: Show only specific providers
  • Language: Narrow to language requirements
  • Gender: Male, female, or gender-neutral
Preview:
  • Click play button on ElevenLabs voices to hear samples
  • Deploy to test agent for extended previews with real scenarios
Metadata displayed:
  • Provider and voice generation technology
  • Language support and multilingual capabilities
  • EU hosting badge
  • Gender, accent, and tone characteristics

​
Next Steps

Voice Settings

Fine-tune speed, pitch, and stability for your voice

Custom Pronunciations

Correct pronunciation of brand names and terms

Transcriber

Configure speech recognition

Ambient Sound

Add background ambience