Handle e-commerce WhatsApp support with Whapi, WooCommerce, RAG and Gmail escalation

This workflow is a complete AI-powered customer support automation for e-commerce businesses.

The system allows customers to interact with an AI assistant directly through WhatsApp. This chatbot can:

• Answer customer questions in real time
• Retrieve order information from WooCommerce
• Search products and provide recommendations
• Access company policies and FAQs using RAG (Retrieval-Augmented Generation)
• Escalate conversations to human support when necessary
• Maintain conversation memory for contextual interactions
• Apply AI guardrails for safer conversations

---

Key Advantages

1. ✅ Fully Automated Customer Support

The workflow dramatically reduces manual customer support workload by automating:

• FAQs
• Order tracking
• Product inquiries
• Return policy questions
• Technical troubleshooting

This enables 24/7 support availability without human intervention.

2. ✅ Native WhatsApp Integration

Using Whapi, customers can communicate through WhatsApp — one of the most widely used messaging platforms globally.

3. ✅ AI-Powered Product Assistance

The assistant can:

• Search products in WooCommerce
• Compare items
• Recommend products
• Check stock availability
• Provide real-time pricing

This transforms the chatbot into an intelligent sales assistant, increasing conversion opportunities.

4. ✅ RAG Knowledge Base with Qdrant

The workflow uses Retrieval-Augmented Generation (RAG) powered by Qdrant vector search.

5. ✅ Human Escalation System

When the AI cannot solve an issue, the workflow automatically escalates the conversation to human support via Gmail.

6. ✅ Multi-Model AI Architecture

The workflow combines:

• Google Gemini for conversational reasoning
• OpenAI embeddings for semantic retrieval

7. ✅ Persistent Conversation Memory

The Window Buffer Memory node maintains contextual conversations.

8. ✅ Built-In Guardrails & Security

The workflow includes AI Guardrails to:

• Filter inappropriate content
• Block unsafe requests
• Prevent policy violations
• Improve AI reliability

This is essential for production-ready AI assistants.

9. ✅Real-Time WooCommerce Integration

The chatbot interacts directly with WooCommerce APIs to:

• Retrieve customer profiles
• Access order history
• Check shipment status
• Query products

This provides live data instead of static responses.

10. ✅ Modular & Scalable Architecture

The workflow is highly modular:

• Easily customizable
• Extendable with new tools
• Compatible with additional APIs
• Scalable for large e-commerce operations

New AI tools or integrations can be added without redesigning the entire system.

Ideal Use Cases

This workflow is ideal for:

• E-commerce stores
• Fashion brands
• Electronics retailers
• Customer support automation
• AI sales assistants
• WhatsApp commerce
• Technical support automation
• Multilingual support systems

This workflow represents a production-ready AI customer support system capable of combining conversational AI, semantic search, real-time e-commerce operations, and human escalation into a single automated pipeline.

---

How it works

1. WhatsApp message reception

• The workflow starts when a WhatsApp message is received via the Get WhatsApp webhook (Whapi).
• It checks if the message is of type text — if not, it replies with a message saying only text is supported.

1. Guardrails & policy check

• The message passes through a Guardrails node that filters inappropriate content or policy violations.
• If the message violates policies, a rejection reply is sent.

1. Customer support AI Agent

• The message is passed to an AI Agent (powered by Google Gemini or OpenAI).
• The agent has access to:
• Memory (Window Buffer Memory) to keep conversation context.
• Tools for:
• Retrieving company knowledge via rag_search (from Qdrant).
• Fetching WooCommerce orders, products, and customers.
• Performing calculations.
• Escalating to human support via Gmail.
• The agent follows a detailed system prompt covering product info, order assistance, technical support, and escalation logic.

1. Response normalization & sending

• The agent's raw output is cleaned (newlines/tabs escaped) via a Code node.
• The cleaned response is sent back to the customer via Whapi (Send WhatsApp).

1. Knowledge base vectorization (separate flow)

• A separate part of the workflow (triggered manually) loads documents from Google Drive, splits them into chunks, generates embeddings via OpenAI, and stores them in Qdrant for semantic search.

---

Set up steps

1. Configure external services & credentials

• Whapi: Sign up (FREE TRIAL available), get API credentials, and set the webhook URL in Whapi settings to point to your n8n webhook.
• OpenAI: Create an API key for embeddings and chat model (or use Google Gemini).
• Qdrant: Set up a Qdrant instance (e.g., on Hetzner) and create API credentials.
• Google Drive: Authenticate to access document folders.
• WooCommerce: Generate API keys for order/product/customer retrieval.
• Gmail: Authenticate for human support escalation.

1. Create Qdrant collection

• Run the Create collection HTTP request node (update QDRANTURL and COLLECTION name).

1. Upload & vectorize documents

• Place your knowledge base files (PDFs, text, Google Docs) in a Google Drive folder.
• Set the folder ID in the Get folder node.
• Run the manual trigger to load, split, embed, and store documents into Qdrant.

1. Configure the AI Agent

• Set the system prompt (already provided in the E-Commerce Customer Support AI Agent node).
• Choose the language model (Gemini or OpenAI) and connect credentials.
• Ensure all tools (rag_search, WooCommerce nodes, calculator, Gmail) are enabled and properly connected.

1. Set webhook & activate workflow

• Copy the webhook URL from the Get WhatsApp node.
• Paste it into your Whapi dashboard as the incoming message webhook.
• Toggle the workflow to active.

---

👉 Subscribe to my new YouTube channel. Here I’ll share videos and Shorts with practical tutorials and FREE templates for n8n.

---

Need help customizing?

Contact me for consulting and support or add me on Linkedin.