WhatsApp Business GPT-4o RAG n8n Workflow w/ Qdrant & Drive

Turn WhatsApp messages into GPT-powered answers using RAG + Qdrant + Google Drive

This n8n workflow connects Meta WhatsApp Business webhooks to an OpenAI-powered RAG assistant that retrieves answers from a Qdrant vector database built from Google Drive documents—then replies directly to your customer in WhatsApp.

What this workflow does

• Webhook verification (GET): Receives the Meta webhook verification request and returns the hub.challenge value.
• Incoming message handling (POST): Listens for WhatsApp events and checks whether the payload contains a user message.
• Text-only support: If the incoming event is not a supported text message, the workflow sends a WhatsApp reply indicating that only text messages are supported.
• RAG-enabled answering: For valid text messages, it extracts the user query and sends it to an OpenAI chat model configured with a business support system prompt.
• Embeddings + retrieval: Generates OpenAI embeddings for the query, retrieves relevant context from a Qdrant collection, and provides the retrieved documents to the agent.
• WhatsApp response: Sends the final agent answer back to the customer using the WhatsApp Business API.

Use cases

• SaaS support: Answer common questions from your knowledge base stored in Google Drive.
• Customer onboarding: Guide users based on product docs, policies, and FAQs indexed in Qdrant.
• Operations teams: Provide consistent WhatsApp support without manually searching documents.

Technical details

• n8n nodes/logic: webhook, if, no op, WhatsApp, sticky note, Google Drive.
• Integrations: Meta WhatsApp Business webhooks, OpenAI (chat + embeddings), Qdrant (vector retrieval), Google Drive (document source).
• Key requirements: Configure WhatsApp callback URL (GET + POST), add OpenAI credentials for chat completions and embeddings, and set up Qdrant credentials and a collection for retrieval.