AI Voice Commerce & Conversational Assistant
Configure multi-provider speech-to-text transcription, conversational product search, automated smart upsells, and direct voice commands.
How Voice Commerce Works
When a shopper sends an audio or voice recording to the Telegram bot:
- The audio file is downloaded securely via the Telegram Bot API.
- Audio is transcribed to text using the configured provider (OpenAI Whisper or Google Gemini).
- The transcript is echoed to the user for immediate visual confirmation.
- VoiceCommandHandler parses store commands (e.g. "What is the status of order #1042?" or "Clear my cart") or executes semantic catalog search via ChatAssistant.
Supported AI Providers
OmniCommerce AI includes out-of-the-box support for leading LLM and Speech providers:
- Google Gemini: Fast multimodal audio processing via
gemini-1.5-flash. - OpenAI: Whisper-1 for speech transcription and GPT-4o-mini for conversational intelligence.
- Groq: Ultra-low latency Llama-3.3-70b inference.
- OpenRouter & NVIDIA NIM: Broad model selection across hundreds of open and closed weights.
- EthioNexus SaaS Proxy: Zero-setup managed AI endpoint (no external provider account required).
Native Voice Commands Reference
Example Commands
"Show me running shoes under 80 dollars" -> Filters catalog by category and price
"Add the blue jacket size medium to cart" -> Matches variable product & updates cart
"What is the status of my last order?" -> Fetches latest order status linked to chat ID
"Clear my cart" -> Empties active cart session
"Help me speak with a human support agent" -> Triggers Live Support handoff
Smart AI Upsell & Cross-Sell Engine
When enabled in OmniCommerce AI → AI Assistant & Voice, the UpsellEngine analyzes cart contents and crafts persuasive, customized upsell pitches directly within notification messages and checkout summaries.