LLM Customer Service Bot for Restaurant Inquiries and Reservations

This workflow automates customer service for restaurants by replacing manual WhatsApp responses to inquiries and bookings, which typically consume 20+ hours weekly for staff handling 100+ messages. It utilizes AI to process queries instantly, reducing response times from hours to seconds and cutting labor costs by up to 70% for small to medium-sized eateries. Key nodes include the WhatsAppTrigger for receiving messages, IF node (DevHubConnect Message Validator) to filter valid texts, a Langchain Agent (DevHubConnect Restaurant AI Agent) with the Ollama model for intelligent replies, a Set node (DevHubConnect Response Processor) to format data, another IF (DevHubConnect Booking Intent Detector) for routing bookings, Postgres nodes for logging and reservations, and WhatsApp nodes for sending responses. This helps restaurant owners and small teams (1-5 staff) manage high-volume customer interactions without hiring extra support.\n\nExpect ROI like saving 15 hours/week on 200+ queries, enabling 24/7 service that boosts customer satisfaction by 40% and increases bookings by 25%. Ideal for hospitality businesses with 50-500 daily patrons; scales to 1,000 messages/month. Requires WhatsApp Business API ($0.005/message via Meta), Ollama (free local AI), and Postgres (free tier on Supabase) subscriptions. n8n Cloud starts at $20/month. Integrates via APIs; limits at WhatsApp's 1,000/day tier, scalable with upgrades.\n\nInstall n8n via cloud.n8n.io signup or self-host from n8n.io (Docker: docker run -it --rm --name n8n -p 5678:5678 n8nio/n8n). For WhatsApp, get API credentials from Meta Business Suite: create an app, add a phone number, and generate an access token (permanent for production). Set in n8n credentials: WhatsAppTriggerApi with token and phone ID (e.g., 550325331503475). Configure webhook URL in Meta dashboard to n8n's /webhook/ endpoint (e.g., https://your-n8n-instance.app/webhook/62614db0-... ). For Ollama, install locally (ollama.ai), run 'ollama serve', and add a credential with base URL http://localhost:11434. Postgres: create DB on Supabase, add tables devhubconnect_reservations and _conversations, input connection string. Update systemMessage in Agent node with restaurant details.\n\nTest by sending sample WhatsApp texts like 'Reserve table for 4 tomorrow' to trigger flow; verify AI response and DB insert. Common errors: invalid token (401, refresh in Meta), rate limits (429, add delays in n8n), webhook mismatch (update URL). Activate via n8n toggle; monitor executions in dashboard. Maintain by weekly DB backups, update Ollama model. Optimize: lower temperature to 0.2 for consistent replies, add error node for alerts.", "businessValue": "Saves 15 hours/week on 200+ customer queries, boosting bookings by 25% automatically", "setupTime": "45-60 minutes", "difficulty": "Intermediate", "requirements": ["WhatsApp Business API subscription ($0.005/message)", "Ollama free local installation", "Postgres DB (free tier on Supabase)", "n8n cloud or self-host ($20+/month)", "Meta developer account for API tokens"], "useCase": "Automating restaurant customer service and reservation bookings via WhatsApp for small hospitality businesses"