Skip to main content

Voice & Language Settings

Customize how your AI assistant sounds and which languages it understands. The right voice and language settings create a natural, professional experience for your customers.

Video Tutorial Available

Video ID: VIDEO-VOICE-001 Title: Customizing Voice & Language Settings Duration: 6-8 minutes Description: Complete guide to selecting voices, languages, and timing settings Status: 🔴 Not recorded yet

Overview​

Your assistant's voice and language settings control two main things:

  1. Voice (Text-to-Speech) - How your assistant speaks to customers
  2. Language (Speech-to-Text) - How your assistant understands what customers say
Quick Tip

You can change these settings anytime without affecting your phone number or integrations. Test different voices to find the perfect fit for your brand!

Choosing a Voice​

Where to Find Voice Settings​

  1. Log in to voice.vokaai.com
  2. Go to AI Assistants
  3. Select the assistant you want to customize
  4. Click the Voice Tab
  5. Find the Voice dropdown
Screenshot Needed

Screenshot ID: VOICE-001 Filename: VOICE-001_voice-tab-location.png Description: Assistant editor with Voice tab highlighted and voice dropdown visible Status: 🔴 Not started

Available Voice Options​

Voka AI offers multiple voice providers with different characteristics:

Best for: Most business use cases

Characteristics:

  • High-quality, natural-sounding voices
  • Low latency (fast responses)
  • Multiple accents and genders
  • Great balance of quality and speed

Popular Voices:

  • VokaAI.en-US-EmmaNeural - Female, American accent, warm and professional
  • VokaAI.en-US-AndrewNeural - Male, American accent, clear and confident
  • VokaAI.en-US-JennyNeural - Female, American accent, friendly and approachable
  • VokaAI.en-CA-ClaraNeural - Female, Canadian accent, polite and clear

When to use:

  • Customer service
  • Appointment scheduling
  • General business communications
  • After-hours coverage

Voka AI Premium​

Best for: Premium brand experiences

Characteristics:

  • Ultra-natural, human-like quality
  • Excellent expressiveness and emotion
  • Slightly higher latency (minimal delay)
  • Perfect for customer-facing roles

Popular Voices:

  • VokaAI.Premium-Emma - Female, conversational and warm
  • VokaAI.Premium-Marcus - Male, authoritative yet friendly
  • VokaAI.Premium-Sophia - Female, elegant and professional

When to use:

  • High-end services (luxury brands, executive offices)
  • Scenarios where voice quality is critical
  • Customer retention calls
  • VIP customer interactions
Premium Pricing

Premium voices may incur slightly higher per-minute costs. See Pricing for details.

Standard Voices​

Best for: Budget-conscious implementations

Characteristics:

  • Clear and understandable
  • Fast response times
  • Lower cost per minute
  • Slightly more robotic than Natural/Premium

When to use:

  • Internal testing
  • High-volume call centers (cost optimization)
  • Simple information delivery
  • IVR menu navigation

Voice Selection by Industry​

Different industries benefit from different voice characteristics:

IndustryRecommended VoiceWhy
HealthcareVokaAI.en-US-EmmaNeuralWarm, reassuring, professional
LegalVokaAI.Premium-MarcusAuthoritative, trustworthy
Beauty/WellnessVokaAI.en-US-JennyNeuralFriendly, upbeat, welcoming
Home ServicesVokaAI.en-US-AndrewNeuralClear, confident, helpful
RestaurantsVokaAI.en-US-JennyNeuralFriendly, fast-paced
Real EstateVokaAI.Premium-SophiaProfessional, polished
FitnessVokaAI.en-US-JennyNeuralEnergetic, motivating

Testing Voices​

Before selecting a voice:

  1. Preview in Dashboard
    • In the Voice tab, click Preview next to any voice
    • Type a sample message your assistant might say
    • Click Play to hear the voice
Screenshot Needed

Screenshot ID: VOICE-002 Filename: VOICE-002_voice-preview.png Description: Voice selection interface with voice preview/play buttons Status: 🔴 Not started

  1. Make Test Calls

    • Save your voice selection
    • Call your assistant using the Test Call feature
    • Evaluate the voice in a real conversation

  2. Get Feedback

    • Have team members test the assistant
    • Consider asking customers (if already live)
    • Trust your instincts - if it sounds right, it probably is!
Pro Tip

The same voice can sound different depending on your instructions. A warm, friendly greeting can make any voice feel more welcoming!

Language Settings​

Supported Languages​

Voka AI supports conversations in 60+ languages including:

Most Common:

  • English (US, UK, Canada, Australia)
  • Spanish (Spain, Mexico, Latin America)
  • French (France, Canada)
  • German
  • Italian
  • Portuguese (Brazil, Portugal)
  • Chinese (Mandarin, Cantonese)
  • Japanese
  • Korean
  • Hindi
  • Arabic

See full list: All Supported Languages

Setting Your Language​

For English-only assistants (most common):

  1. Go to Voice Tab in your assistant settings
  2. Find Language setting
  3. Select English (en-US) or your regional variant
  4. Save

For multi-language support:

  • Select Multi-Language Detection
  • Your assistant will automatically detect and respond in 60+ languages
  • Best for: Diverse customer bases, tourism, international businesses
Language vs. Voice

Make sure your Voice and Language match! An English voice with Spanish language detection won't work well. Choose:

  • English voice → English language
  • Spanish voice → Spanish language
  • Or use Multi-Language with a neutral accent

How Language Detection Works​

Single Language Mode:

  • You specify the language (e.g., English)
  • Assistant only understands that language
  • Faster, more accurate for that specific language
  • Recommended for: Businesses with customers speaking one language

Multi-Language Mode:

  • Assistant automatically detects the language being spoken
  • Responds in the detected language (if supported)
  • Slightly higher processing time
  • Recommended for: International businesses, tourism, multilingual communities

Advanced Voice Settings​

Speech Rate​

Control how fast your assistant speaks.

Settings:

  • Slow (0.75x speed) - Clearer for complex information
  • Normal (1.0x speed) - Natural conversational pace (recommended)
  • Fast (1.25x speed) - Quick responses, efficient
Screenshot Needed

Screenshot ID: VOICE-003 Filename: VOICE-003_speech-rate-slider.png Description: Advanced settings showing speech rate slider control Status: 🔴 Not started

When to adjust:

  • Slower: Medical info, legal disclaimers, elderly customers
  • Faster: Quick confirmations, high-volume call centers

How to configure:

  1. Voice Tab → Advanced Settings
  2. Adjust Speech Rate slider
  3. Test with preview
  4. Save

Pitch Adjustment​

Fine-tune the voice's pitch (tone).

Settings:

  • Lower - More authoritative, masculine
  • Normal - Natural voice tone (recommended)
  • Higher - Brighter, more energetic

When to adjust:

  • Match your brand personality
  • Differentiate between multiple assistants
  • Create specific character for role-play scenarios
Subtle Changes Work Best

Small adjustments (±10%) sound most natural. Large pitch changes can sound robotic.

Pronunciation Customization​

Teach your assistant to pronounce specific words correctly.

Common use cases:

  • Business name (e.g., "Voka AI" pronounced "Voe-kuh A-I")
  • Industry jargon
  • Product names
  • Local place names
  • Owner/staff names

How to configure:

  1. Go to Voice Tab → Pronunciation
  2. Click Add Custom Pronunciation
  3. Enter:
    • Word: The text as written (e.g., "Voka")
    • Phonetic: How to say it (e.g., "VOE-kuh")
  4. Save

Example configurations:

Word/PhrasePhonetic SpellingUse Case
Voka AIVOE-kuh A-ICompany name
Acuityah-KYOO-ih-teeSoftware name
HVACH-V-A-CIndustry term
St. Louissaint LOO-issCity name
Use the Preview

After adding pronunciations, preview your assistant saying a sentence with the custom word to verify it sounds correct.

Response Timing Settings​

Control when your assistant starts speaking after a customer finishes.

Why Timing Matters​

Too fast:

  • Interrupts customers mid-sentence
  • Feels robotic and rushed
  • Customers get frustrated

Too slow:

  • Awkward pauses
  • Customers repeat themselves
  • Conversation feels unnatural

Just right:

  • Natural conversation flow
  • Customers feel heard
  • Professional experience

Timing Options​

Wait Time (Default: 0.4 seconds)​

Purpose: How long to wait after customer stops talking

Recommended settings:

  • 0.3s - Quick, snappy responses (customer service)
  • 0.4s - Standard, natural pace (recommended)
  • 1.0s+ - Patient, thoughtful responses (complex scenarios)

How to configure:

  1. Voice Tab → Advanced Settings
  2. Adjust Response Wait Time slider
  3. Test with different scenarios
  4. Save

Interruption Sensitivity​

Purpose: How aggressive your assistant is about interrupting

Settings:

  • Low - Rarely interrupts, waits for clear pauses
  • Medium - Balanced (recommended)
  • High - Jumps in quickly when there's a pause

When to adjust:

SettingBest For
LowCustomers who think while speaking, elderly, complex info gathering
MediumGeneral business, most use cases
HighQuick transactions, confirmations, IVR navigation
Be Careful with High Sensitivity

High interruption sensitivity can frustrate customers who pause to think. Test thoroughly!

Special Timing: Numbers & Phone Numbers​

When customers read numbers (phone numbers, account numbers), they often pause between digits. Configure special handling:

Number Wait Time (Default: 0.5s)

  • 0.5s - Standard for most number reading
  • 1.0s - Elderly customers or complex numbers
  • 0.3s - Quick confirmations

Example scenario:

Customer: "My phone number is... 555... 0123"

With proper number timing: ✅ Assistant waits patiently

Without: ❌ "I didn't catch your—" (interrupts mid-number)

How to configure:

  1. Voice Tab → Advanced Settings → Timing
  2. Adjust Number Reading Wait Time
  3. Save

Background Audio​

Add subtle background sounds to make conversations more natural and human-like.

Why Use Background Audio?​

Benefits:

  • Makes AI sound more human
  • Covers awkward silences during processing
  • Creates atmosphere (office, restaurant, etc.)
  • Keeps customers engaged

When it helps most:

  • Longer tool calls (checking availability, processing bookings)
  • AI thinking/processing time
  • Natural conversation pauses

Background Audio Options​

Preset Options:

  • Office - Subtle keyboard typing, distant conversations
  • Call Center - Professional office environment
  • None - Complete silence (default)

Custom Audio:

  • Upload your own .mp3 file
  • Must be publicly accessible URL
  • Recommended: Subtle, non-distracting ambient sounds

How to configure:

  1. Voice Tab → Background Audio
  2. Select from dropdown or enter custom URL
  3. Preview to hear the effect
  4. Save
Keep It Subtle

Background audio should be barely noticeable - just enough to make the call feel more natural. Loud or distracting backgrounds frustrate customers.

Multi-Language Assistants​

Setting Up Multi-Language Support​

For businesses serving diverse customers:

Step 1: Enable Multi-Language

  1. Voice Tab → Language
  2. Select Auto-Detect (60+ languages)
  3. Save

Step 2: Choose Appropriate Voice

  • Use a neutral accent (reduces confusion)
  • Consider VokaAI.en-US-EmmaNeural or similar
  • Voice will adapt pronunciation to detected language

Step 3: Update Instructions

  • Include multi-language context in your assistant's instructions
  • Example: "You can communicate in English, Spanish, French, or any of 60+ languages. Detect the customer's language and respond accordingly."

Step 4: Test Multiple Languages

  • Make test calls in different languages
  • Verify detection accuracy
  • Check response quality in each language

Multi-Language Best Practices​

Do's: ✅ Keep instructions clear and simple (easier to translate) ✅ Test with native speakers of each target language ✅ Consider cultural differences in greetings ✅ Be patient with automatic detection (it's fast but not instant)

Don'ts: ❌ Use highly idiomatic English in instructions (doesn't translate well) ❌ Assume voice quality is equal across all languages (test first) ❌ Forget to update knowledge base for multiple languages ❌ Mix languages within the same conversation (confuses customers)

Voice Quality Troubleshooting​

"Voice sounds robotic"​

Possible causes:

  1. Using Standard voices instead of Natural/Premium
  2. Speech rate set too fast or slow
  3. Pronunciation issues with key words
  4. Instructions too formal or stiff

Fixes:

  • Upgrade to VokaAI Natural or Premium voices
  • Adjust speech rate to 1.0x (normal)
  • Add custom pronunciations for proper nouns
  • Rewrite instructions to be more conversational
  • See Prompts Guide for tips

"Assistant interrupts customers"​

Possible causes:

  1. Response wait time too short
  2. Interruption sensitivity too high
  3. Number timing not configured

Fixes:

  • Increase response wait time to 0.5s or higher
  • Lower interruption sensitivity
  • Configure number reading timing to 1.0s
  • Test with real customers and adjust

"Voice doesn't match brand"​

Solution:

  • Try different voices (preview 5-10 options)
  • Adjust pitch slightly (±10%)
  • Consider Premium voices for high-end brands
  • Modify instructions to reflect brand personality
  • Get team feedback on voice selection

"Language detection is wrong"​

Possible causes:

  1. Customer has strong accent
  2. Multi-language mode when single language would be better
  3. Background noise interfering

Fixes:

  • If customers all speak one language, use single-language mode
  • Improve call quality (reduce background noise)
  • Use higher quality phone lines/numbers
  • Consider manual language selection in greeting

"Voice cuts out or stutters"​

Possible causes:

  1. Network connectivity issues
  2. Using Premium voices with poor internet
  3. Complex pronunciations processing slowly

Fixes:

  • Check your internet connection stability
  • Switch to VokaAI Natural (lower latency)
  • Simplify pronunciation customizations
  • Contact support if persistent

Complete Language List​

Fully Supported Languages (60+)​

Americas:

  • English (US, Canada, Australia, UK, India)
  • Spanish (Spain, Mexico, Argentina, Colombia)
  • Portuguese (Brazil, Portugal)
  • French (France, Canada)

Europe:

  • German, Italian, Dutch, Polish
  • Russian, Ukrainian
  • Swedish, Norwegian, Danish, Finnish
  • Czech, Romanian, Hungarian
  • Greek, Turkish

Asia:

  • Chinese (Mandarin, Cantonese)
  • Japanese, Korean
  • Hindi, Tamil, Telugu
  • Thai, Vietnamese
  • Indonesian, Malay
  • Arabic (Modern Standard, Egyptian, Gulf)

Others:

  • Hebrew
  • Afrikaans
  • Swahili
  • And 30+ more regional variants

For a complete, up-to-date list: Contact support

Voice Samples & Demos​

Listen to Voice Samples​

In the Dashboard:

  1. Go to any assistant → Voice Tab
  2. Click Preview next to any voice
  3. Enter sample text: "Hello! Thanks for calling. How can I help you today?"
  4. Click Play

Compare Multiple Voices:

  • Open voices in separate browser tabs
  • Test same message in each voice
  • Note differences in tone, speed, clarity
  • Save your favorites

Best Practices Summary​

Choosing a Voice: ✅ Test 5-10 voices before deciding ✅ Match voice to your industry and brand ✅ Get feedback from team and customers ✅ Consider gender diversity (have male and female options)

Language Settings: ✅ Use single language mode if possible (better accuracy) ✅ Match voice language to customer base ✅ Test multi-language thoroughly before going live ✅ Update knowledge base for all supported languages

Timing & Responsiveness: ✅ Start with default settings (0.4s wait time) ✅ Adjust based on real customer feedback ✅ Lower sensitivity for elderly/thoughtful customers ✅ Test extensively with real use cases

Background Audio: ✅ Keep it subtle (barely noticeable) ✅ Match to your business (office for office, none for medical) ✅ Test that it doesn't distract from conversation

Pronunciation: ✅ Add custom pronunciations for your business name ✅ Include industry-specific terms ✅ Test pronunciation changes with preview ✅ Limit to words that really need it (< 20 words)

Next Steps​

Now that you've customized your assistant's voice and language:

  1. Write better instructions - Improve what your assistant says
  2. Test thoroughly - Verify voice sounds great in real scenarios
  3. Add tools - Enable booking, transfers, etc.
  4. Set up knowledge base - Give your assistant information to share

Need Help?​

Questions about voice settings?