AI Profiles Module Documentation
Table of Contents
Section titled “Table of Contents”- Navigation & Access
- Screenshots & Visual Interface
- 🎯 User Roles & Key Capabilities
- Module Overview (Technical)
- Module Overview (Commercial/Business)
- Module Overview (End User/Administrator)
- Configuration Sections
- Settings Reference
- Model Context Protocol (MCP) AI Integration
- Common Scenarios & Examples
- Limitations & Important Notes
- Troubleshooting Tips
- Glossary
Navigation & Access
Section titled “Navigation & Access”To access the AI Profiles module:
- Log in to the Ring2All Web Portal (
https://<domain-or-ip>/login). - In the left navigation sidebar, expand Administration.
- Under AI Integration, click AI Profiles (
/admin/ai-integration/profiles). - Review available AI profiles, bound provider engines, model families (GPT-4o, Claude 3.5, Whisper, Llama 3.3), temperature parameters, and active statuses.
- Click + Add to construct a new specialized model profile (
/admin/ai-integration/profiles/new), or edit existing profile hyperparameters.
Screenshots & Visual Interface
Section titled “Screenshots & Visual Interface”AI Profiles Directory
Section titled “AI Profiles Directory”Central catalog displaying configured AI runtime profiles, bound provider engines, target models, token limits, temperature values, and service scopes.
Create / Edit AI Profile Configuration Form
Section titled “Create / Edit AI Profile Configuration Form”Profile specification editor for naming profiles, choosing providers, selecting target models, tuning temperature (creativity vs determinism), configuring maximum token generation limits, and assigning service scopes.
🎯 User Roles & Key Capabilities
Section titled “🎯 User Roles & Key Capabilities”| Role | Access Level | Responsibilities & Capabilities |
|---|---|---|
| PBX Super Administrator | Full Access (RW) |
Create and calibrate multi-modal AI profiles (LLM, STT, TTS, Embeddings), assign provider engines, configure temperature, and set system prompts. |
| AI Prompt & Voice Engineer | Profile Tuning (RW) |
Tune conversational parameters (temperature, max tokens), configure Google TTS / ElevenLabs voice prosody (pitch, rate), and test sample audio synthesis. |
| Contact Center Supervisor | Reusable Selection (RO) |
Inspect which AI Profiles power inbound call transcription, sentiment analysis, agent copilot assistance, and real-time speech translation. |
| AI Platform Copilot / MCP Agent | Programmatic Audit (RO) |
Execute list_ai_profiles and get_ai_profile_status to evaluate model parameters and recommend appropriate profiles for telephony workflows. |
1. Module Overview (Technical)
Section titled “1. Module Overview (Technical)”What Are AI Profiles?
Section titled “What Are AI Profiles?”AI Profiles is an AI configuration module that defines how AI services behave in the PBX system. Each profile configures the AI provider, model settings, voice parameters, and behavior for features like TTS, transcription, and intelligent assistants.
Architecture
Section titled “Architecture”┌─────────────────────────────────────────────────────────────────┐│ AI Profiles Architecture │├─────────────────────────────────────────────────────────────────┤│ ││ AI Profile: Customer Service ││ ┌──────────────────────────────────────────────────────────┐ ││ │ │ ││ │ API Connection: │ ││ │ ├─ Provider: OpenAI │ ││ │ └─ Model: GPT-4o │ ││ │ │ ││ │ Model Behavior: │ ││ │ ├─ Temperature: 0.7 (creative) │ ││ │ ├─ Max Tokens: 2000 │ ││ │ └─ Embeddings: text-embedding-3-small │ ││ │ │ ││ │ Voice Settings: │ ││ │ ├─ Voice Engine: OpenAI TTS │ ││ │ ├─ Voice: Alloy │ ││ │ ├─ Pitch: Medium │ ││ │ ├─ Volume: Medium │ ││ │ └─ Speech Rate: Normal │ ││ │ │ ││ │ System Prompt: "You are a helpful customer service..." │ ││ │ │ ││ └──────────────────────────────────────────────────────────┘ ││ │ ││ ▼ Referenced by AI Features ││ ┌──────────────────────────────────────────────────────────┐ ││ │ │ ││ │ Recording TTS ──────────► Profile: Studio Voice │ ││ │ AI Assistant ───────────► Profile: Customer Service │ ││ │ Voicemail Summary ──────► Profile: Analysis │ ││ │ IVR Agent ──────────────► Profile: IVR Bot │ ││ │ │ ││ └──────────────────────────────────────────────────────────┘ ││ │└─────────────────────────────────────────────────────────────────┘2. Module Overview (Commercial/Business)
Section titled “2. Module Overview (Commercial/Business)”Business Value
Section titled “Business Value”AI Profiles provides customized AI behavior:
| Without Profiles | With Profiles |
|---|---|
| Single AI config | Multiple personalities |
| Fixed voice | Customizable TTS |
| No tuning | Temperature, tokens |
| One-size-fits-all | Task-specific AI |
Use Cases
Section titled “Use Cases”-
Customer Service Bot
- Friendly, helpful tone
- Moderate creativity
-
Professional Announcements
- Clear, formal voice
- Low creativity (consistent)
-
Call Transcription
- Accurate model
- Fast processing
-
IVR Assistant
- Natural voice
- Specific system prompt
Feature Highlights
Section titled “Feature Highlights”| Feature | Benefit |
|---|---|
| Multiple Profiles | Different use cases |
| Voice Customization | Pitch, rate, volume |
| Temperature Control | Creativity level |
| System Prompts | Define AI behavior |
| TTS Testing | Preview voices |
| Embeddings | Semantic search |
3. Module Overview (End User/Administrator)
Section titled “3. Module Overview (End User/Administrator)”What Can You Do?
Section titled “What Can You Do?”- Create AI profiles for different purposes
- Select providers and models
- Configure temperature and tokens
- Choose TTS voice and settings
- Write system prompts
- Test voice generation
- Adjust pitch, volume, rate
AI Profiles - List View
Section titled “AI Profiles - List View”┌─────────────────────────────────────────────────────────────────┐│ AI Profiles │├─────────────────────────────────────────────────────────────────┤│ ││ Configure AI integration settings, voice profiles, and ││ intelligent assistants ││ ││ [+ Add] ││ ││ ┌───────────────────────────────────────────────────────────┐ ││ │ Name │ Provider │ Model │ Voice │ Status│ ││ ├──────────────────┼──────────┼─────────┼──────────┼───────┤ ││ │ Customer Service │ OpenAI │ GPT-4o │ Alloy │ Active│ ││ │ Studio Voice │ OpenAI │ - │ Nova │ Active│ ││ │ Analysis │ Anthropic│ Claude 3│ - │ Active│ ││ │ IVR Bot │ OpenAI │ GPT-4 │ Shimmer │ Active│ ││ └───────────────────────────────────────────────────────────┘ ││ ││ Actions: [🧪 Test] [✏️ Edit] [🗑️ Delete] ││ │└─────────────────────────────────────────────────────────────────┘Add/Edit AI Profile
Section titled “Add/Edit AI Profile”┌─────────────────────────────────────────────────────────────────┐│ Edit AI Profile │├─────────────────────────────────────────────────────────────────┤│ ││ [API Connection] [Model Config] [Voice Settings] [System] ││ ││ ▼ API Connection ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Profile Name: [Customer Service ] │││ │ Unique identifier for this AI profile │││ │ │││ │ Provider: [OpenAI ▼] │││ │ OpenAI | Anthropic | Google | Hugging Face | Local │││ │ │││ │ Model: [GPT-4o ▼] │││ │ Default AI model for reasoning and responses │││ │ │││ │ [Test Connection] │││ │ │││ └─────────────────────────────────────────────────────────────┘││ ││ ▼ Model Behavior ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Temperature: [0.7 ] (0.0 - 1.0) │││ │ Lower = focused, Higher = creative │││ │ │││ │ Max Tokens: [2000 ] (500 - 8000) │││ │ Maximum response length │││ │ │││ │ Embeddings Model: [text-embedding-3-small ▼] │││ │ Model for semantic search and indexing │││ │ │││ └─────────────────────────────────────────────────────────────┘││ ││ ▼ System Settings ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Timeout (ms): [30000 ] (1000 - 60000) │││ │ │││ │ Retry Attempts: [3 ] (0 - 5) │││ │ │││ │ Status: ✓ Enabled │││ │ │││ └─────────────────────────────────────────────────────────────┘││ ││ [Save] [Cancel] ││ │└─────────────────────────────────────────────────────────────────┘Voice Settings Tab
Section titled “Voice Settings Tab”┌─────────────────────────────────────────────────────────────────┐│ Edit AI Profile - Voice Settings │├─────────────────────────────────────────────────────────────────┤│ ││ ▼ Voice Selection ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Voice Engine: [OpenAI TTS ▼] │││ │ OpenAI TTS | ElevenLabs | AWS Polly | Google TTS │││ │ │││ │ Voice: [🔍 Search voices... ] │││ │ Alloy | Echo | Fable | Nova | Onyx | Shimmer │││ │ │││ │ ─── Google TTS Selection (if Google TTS selected) ─── │││ │ Language: [🔍 Search language... ] │││ │ Spanish | English | French | Portuguese | German... │││ │ Accent: [Spain ▼] │││ │ Spain | Mexico | US | Argentina... │││ │ Voice: [🔍 Enceladus ▼] │││ │ Chirp D - Female | Chirp E - Male | Enceladus - Female │││ │ Style: [Natural ▼] │││ │ Natural | Alegre | Triste | Susurrar | Storyteller │││ │ │││ └─────────────────────────────────────────────────────────────┘││ ││ ▼ Voice Configuration ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Pitch: [Medium ▼] │││ │ Very Low | Low | Medium | High | Very High │││ │ │││ │ Volume: [Medium ▼] │││ │ Silent | Very Soft | Soft | Medium | Loud | Very Loud │││ │ │││ │ Speech Rate: [Medium ▼] │││ │ Very Slow | Slow | Medium | Fast | Very Fast │││ │ │││ └─────────────────────────────────────────────────────────────┘││ ││ ▼ Test Voice ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ Text to convert: [Hello, welcome to our service ] │││ │ │││ │ [Generate] [🔊 Play] │││ │ │││ │ [Download Production] │││ │ │││ └─────────────────────────────────────────────────────────────┘││ │└─────────────────────────────────────────────────────────────────┘System Prompt Tab
Section titled “System Prompt Tab”┌─────────────────────────────────────────────────────────────────┐│ Edit AI Profile - System Settings │├─────────────────────────────────────────────────────────────────┤│ ││ ▼ System Prompt ││ ││ ┌─────────────────────────────────────────────────────────────┐││ │ │││ │ System Prompt: │││ │ ┌─────────────────────────────────────────────────────┐ │││ │ │ You are a helpful customer service assistant for │ │││ │ │ a telecommunications company. Be friendly, concise,│ │││ │ │ and helpful. Guide users to solutions efficiently. │ │││ │ │ │ │││ │ │ Key behaviors: │ │││ │ │ - Greet customers warmly │ │││ │ │ - Understand their needs quickly │ │││ │ │ - Provide clear, step-by-step help │ │││ │ │ - Offer to transfer to human agent if needed │ │││ │ └─────────────────────────────────────────────────────┘ │││ │ Instructions that define how the AI should behave │││ │ │││ └─────────────────────────────────────────────────────────────┘││ │└─────────────────────────────────────────────────────────────────┘Quick Tips
Section titled “Quick Tips”[!TIP] Temperature: Use 0.3 for accuracy, 0.7+ for creativity.
[!TIP] Test Voice: Always preview before using in production.
[!NOTE] System Prompt: Define AI personality and behavior.
4. Configuration Sections
Section titled “4. Configuration Sections”API Connection
Section titled “API Connection”| Field | Description |
|---|---|
| Profile Name | Unique identifier |
| Provider | AI service (OpenAI, etc.) |
| Model | LLM model to use |
Model Behavior
Section titled “Model Behavior”| Field | Description |
|---|---|
| Temperature | Creativity (0.0-1.0) |
| Max Tokens | Response length limit |
| Embeddings Model | For semantic search |
Voice Settings
Section titled “Voice Settings”| Field | Description |
|---|---|
| Voice Engine | TTS provider |
| Voice | Voice selection |
| Pitch | Voice pitch level |
| Volume | Output volume |
| Speech Rate | Speaking speed |
System Settings
Section titled “System Settings”| Field | Description |
|---|---|
| Timeout | API response timeout |
| Retry Attempts | Retry on failure |
| Status | Enable/disable |
| System Prompt | AI behavior instructions |
5. Settings Reference
Section titled “5. Settings Reference”Providers
Section titled “Providers”| Provider | Models | Voice Engines |
|---|---|---|
| OpenAI | GPT-4o, GPT-4 Turbo | OpenAI TTS |
| Anthropic | Claude 3 Opus | - |
| Gemini | Google TTS | |
| Local | Local LLM | Local TTS |
Temperature Guide
Section titled “Temperature Guide”| Value | Behavior | Use Case |
|---|---|---|
| 0.0-0.3 | Focused, consistent | Transcription, facts |
| 0.4-0.6 | Balanced | General assistance |
| 0.7-1.0 | Creative, varied | Creative content |
Voice Engines
Section titled “Voice Engines”| Engine | Provider | Features |
|---|---|---|
| OpenAI TTS | OpenAI | High quality, fast |
| ElevenLabs | ElevenLabs | Voice cloning |
| AWS Polly | Amazon | Neural voices |
| Google TTS | 300+ voices, cascading selection, emotion styles |
Google TTS Selection Hierarchy
Section titled “Google TTS Selection Hierarchy”Google TTS uses a 4-stage cascading selection to manage 300+ voice options:
| Stage | Description | Example |
|---|---|---|
| Language | Base language | Spanish, English |
| Accent | Regional variant | Spain, Mexico, US |
| Voice | Technical variant with persona | Enceladus, Chirp D |
| Style | Emotional tone | Natural, Alegre, Triste |
Google TTS Voice Types
Section titled “Google TTS Voice Types”| Type | Quality | Cost | Use Case |
|---|---|---|---|
| Standard | Basic | Low | Testing, drafts |
| Wavenet | High | Medium | Production |
| Neural2 | Premium | High | Professional |
| Chirp3-HD | Ultra | Premium | Studio quality |
Google TTS Prosody Controls
Section titled “Google TTS Prosody Controls”| Control | Range | Default | Effect |
|---|---|---|---|
| Speed | 0.25x - 4.0x | 1.0x | Speaking rate |
| Pitch | -20 to +20 semitones | 0 | Voice tone |
| Volume | -10 to +10 dB | 0 dB | Output level |
Pitch Levels
Section titled “Pitch Levels”| Level | Value | Effect |
|---|---|---|
| Very Low | -50% | Deep voice |
| Low | -25% | Slightly lower |
| Medium | 0% | Normal |
| High | +25% | Slightly higher |
| Very High | +50% | High pitched |
Speech Rate Levels
Section titled “Speech Rate Levels”| Level | Speed | Use Case |
|---|---|---|
| Very Slow | 0.5x | Clear dictation |
| Slow | 0.75x | Instructions |
| Medium | 1.0x | Normal |
| Fast | 1.5x | Quick info |
| Very Fast | 2.0x | Rapid playback |
Model Context Protocol (MCP) AI Integration
Section titled “Model Context Protocol (MCP) AI Integration”The AI Profiles module integrates with the Model Context Protocol (MCP), enabling autonomous agents and the Platform Copilot to discover, inspect, and evaluate model runtime presets across the PBX platform.
Available MCP Tools
Section titled “Available MCP Tools”| Tool Name | Scope | Description |
|---|---|---|
list_ai_profiles |
Profile Registry (RO) |
List all configured AI Model Profiles (type: LLM, STT, TTS, embeddings, model name, and assigned provider). |
get_ai_profile_status |
Hyperparameter Audit (RO) |
Get configuration parameters of an AI Profile (model, temperature, max tokens, system prompt, timeout). |
Tool Schemas & Payloads
Section titled “Tool Schemas & Payloads”list_ai_profiles
Section titled “list_ai_profiles”{ "name": "list_ai_profiles", "description": "List all configured AI Model Profiles (type: LLM, STT speech-to-text, TTS text-to-speech, embeddings, model name, and assigned provider).", "parameters": { "type": "object", "properties": { "type": { "type": "string", "enum": ["llm", "stt", "tts", "embeddings", "general"], "description": "Filter by profile capability type." }, "search": { "type": "string", "description": "Filter by profile name or model." } } }}Realistic Execution Response:
{ "success": true, "data": { "total": 3, "profiles": [ { "id": 1, "name": "GPT-4o Customer Service Assistant", "type": "llm", "model": "gpt-4o", "providerId": 1, "providerName": "Primary OpenAI Account", "temperature": 0.5, "maxTokens": 2048, "isActive": true }, { "id": 2, "name": "Deepgram Nova-2 Fast Transcription", "type": "stt", "model": "nova-2-general", "providerId": 2, "providerName": "Deepgram Speech Cloud", "isActive": true }, { "id": 3, "name": "ElevenLabs Studio Voice Generator", "type": "tts", "model": "eleven_multilingual_v2", "providerId": 3, "providerName": "ElevenLabs Voice Cloud", "voice": "Rachel", "isActive": true } ] }}get_ai_profile_status
Section titled “get_ai_profile_status”{ "name": "get_ai_profile_status", "description": "Get configuration parameters of an AI Profile (model, temperature, max tokens, system prompt, timeout).", "parameters": { "type": "object", "properties": { "profileName": { "type": "string", "description": "Name of the AI Profile (e.g. \"GPT-4o Customer Service Assistant\")." } }, "required": ["profileName"] }}Realistic Execution Response:
{ "success": true, "data": { "profile": { "id": 1, "name": "GPT-4o Customer Service Assistant", "type": "llm", "model": "gpt-4o", "providerName": "Primary OpenAI Account", "temperature": 0.5, "maxTokens": 2048, "systemPrompt": "You are a courteous, professional customer support agent for Ring2All Telephony...", "timeoutSeconds": 30, "presencePenalty": 0.0, "frequencyPenalty": 0.1 } }}Bilingual Natural Language Prompt Examples
Section titled “Bilingual Natural Language Prompt Examples”English Prompts
Section titled “English Prompts”- “Copilot, list all active LLM profiles configured for telephony voice agents.”
- “What model and temperature are assigned to the Customer Service Assistant profile?”
- “Find all speech-to-text (STT) profiles available for real-time call transcription.”
Spanish Prompts
Section titled “Spanish Prompts”- “Copilot, lista todos los perfiles de IA activos y sus modelos asignados.”
- “¿Cuál es la temperatura y el límite de tokens del perfil de atención a clientes?”
- “Muestra los perfiles de síntesis de voz (TTS) configurados en el sistema.”
Enterprise Safeguards & Execution Boundaries
Section titled “Enterprise Safeguards & Execution Boundaries”- Multi-Tenant Profile Isolation: Profiles are bound to numeric
tenant_id(or shared platform default templates). Sub-tenants cannot overwrite core profiles. - Deterministic Hyperparameter Range Checking: Temperature values must strictly fall within
0.0and2.0; max tokens cannot exceed model context windows. - Dependency Protection: Deleting an AI Profile is blocked if it is actively linked to a live Voice Agent, Extension Chatbot, or Service Hub mapping.
6. Common Scenarios & Examples
Section titled “6. Common Scenarios & Examples”Scenario 1: Create Customer Service Profile
Section titled “Scenario 1: Create Customer Service Profile”- Add Profile
- Name = “Customer Service”
- Provider = OpenAI
- Model = GPT-4o
- Temperature = 0.7
- Max Tokens = 2000
- Voice Engine = OpenAI TTS
- Voice = Alloy
- System Prompt = “You are a helpful…”
- Save
Scenario 2: Studio Voice Profile
Section titled “Scenario 2: Studio Voice Profile”- Add Profile
- Name = “Studio Voice”
- Voice Engine = ElevenLabs
- Voice = (select professional voice)
- Pitch = Medium
- Rate = Slow
- Test Voice
- Save
Scenario 3: Analysis Profile
Section titled “Scenario 3: Analysis Profile”- Add Profile
- Name = “Analysis”
- Provider = Anthropic
- Model = Claude 3 Opus
- Temperature = 0.3 (accurate)
- Max Tokens = 4000
- (No voice settings needed)
- Save
Scenario 4: IVR Bot Profile
Section titled “Scenario 4: IVR Bot Profile”- Add Profile
- Name = “IVR Bot”
- Provider = OpenAI
- Model = GPT-4o
- Temperature = 0.5
- Voice = Nova
- Rate = Medium
- System Prompt = “You are an IVR assistant…”
- Test Voice with sample greeting
- Save
Scenario 5: Google TTS Profile (Spanish)
Section titled “Scenario 5: Google TTS Profile (Spanish)”- Add Profile
- Name = “Spanish Voice”
- Provider = Google
- Voice Engine = Google TTS
- Language = Spanish
- Accent = Spain (or Mexico, US, etc.)
- Voice = Enceladus (or select from Chirp, Wavenet, Neural2)
- Style = Natural (or Alegre for friendly, Triste for empathetic)
- Adjust Speed, Pitch, Volume as needed
- Enter test text and click “Convert to Audio”
- Preview, then Save
7. Limitations & Important Notes
Section titled “7. Limitations & Important Notes”Technical Notes
Section titled “Technical Notes”[!NOTE] Provider Required: Must select a provider from AI Providers first.
[!NOTE] Voice Testing: Uses API credits for TTS generation.
[!WARNING] System Prompt: Keep concise for faster responses.
Best Practices
Section titled “Best Practices”- Use Profiles for Purposes: Don’t reuse same profile for everything
- Test Voice Settings: Preview before production
- Tune Temperature: Match to use case
- Write Clear Prompts: Specific instructions work best
- Monitor Usage: API calls cost money
Profile Use Cases
Section titled “Profile Use Cases”| Profile Type | Temperature | Voice | Example |
|---|---|---|---|
| Customer Service | 0.6-0.7 | Friendly | Alloy, Nova |
| Announcements | 0.2-0.3 | Professional | Onyx |
| Analysis | 0.1-0.3 | None | - |
| Creative | 0.7-1.0 | Varied | - |
8. Troubleshooting Tips
Section titled “8. Troubleshooting Tips”Common Issues
Section titled “Common Issues”| Symptom | Possible Cause | Solution |
|---|---|---|
| Connection failed | Wrong provider | Check AI Providers |
| Voice won’t play | Audio blocked | Check browser |
| Slow responses | Low timeout | Increase timeout |
| Generic answers | Vague prompt | Improve system prompt |
Test Connection
Section titled “Test Connection”# Test API manuallycurl https://api.openai.com/v1/chat/completions \ -H "Authorization: Bearer $API_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"gpt-4o","messages":[{"role":"user","content":"Hello"}]}'Debug Voice
Section titled “Debug Voice”- Click Test Voice
- Enter sample text
- Click Generate
- Check for error messages
- Verify voice is selected
- Check audio playback
9. Glossary
Section titled “9. Glossary”| Term | Definition |
|---|---|
| Temperature | AI creativity control |
| Max Tokens | Response length limit |
| TTS | Text-to-Speech |
| Embeddings | Vector representations |
| System Prompt | AI behavior instructions |
| Voice Engine | TTS service provider |
Documentation last updated: January 2026