All Projects
AI AgentsFeatured
WhatsApp AI Sales Agent
Production AI sales agent on WhatsApp Business. Classifies messages (Sales/Support/Off-topic), queries Supabase product DB, uses Ollama/Llama3.1 locally, bilingual FR/AR, conversation memory. -90% manual processing time.
-90%
Processing time reduction
95%+
Classification accuracy
FR / AR
Languages supported
41
n8n nodes
Approach
Intent classification → specialized AI agent routing with database-grounded responses
Tech Stack
n8nOllamaLlama3.1SupabaseWhatsApp Business APIJavaScript
Keywords
n8nLLMWhatsAppOllamaLlama3.1SupabaseAutomation
Deep Dive
Fully automated AI sales agent for a Moroccan e-commerce client spending 6+ hours/day on repetitive WhatsApp product queries.
Architecture
WhatsApp Business API → n8n Webhook
→ Text Classifier (LLM intent detection)
├── SALES → AI Agent (Llama3.1 + db_product tool)
├── SUPPORT → Support Agent
└── OFF_TOPIC → Auto-rejection message
→ Supabase (product DB + conversation memory)
→ WhatsApp Business API (reply)
Why Ollama (Local LLM)?
- ▸No customer data leaves the server (GDPR/privacy compliance)
- ▸Zero API costs for high-volume messaging
- ▸Full control over model behavior and prompt templates
Anti-Hallucination System Prompt
"NEVER answer product questions without calling db_product first. ONLY present data returned by the tool. If product not found → honestly inform the customer."
Results
| Metric | Value |
|---|---|
| Manual processing time reduction | -90% |
| Message classification accuracy | 95%+ |
| Languages supported | FR / AR (auto-detected) |
| n8n workflow nodes | 41 |