📊 Description
📊 Description This workflow helps travel, booking, and customer support businesses create an AI chatbot that answers customer questions from their own knowledge base. Instead of manually replying to repeated questions about bookings, cancellations, refunds, pricing, policies, or service details, the business can upload a knowledge document and let the chatbot respond through a webhook. Built for travel agencies, booking platforms, hotels, shortlets, service businesses, and automation builders who want to create a simple RAG chatbot using Google Drive, OpenAI, Pinecone, and n8n. What This Workflow Does 📄 Downloads a knowledge document from Google Drive ✂️ Splits the document into smaller chunks for better retrieval 🧠 Creates embeddings using OpenAI 📦 Stores the knowledge base inside a Pinecone vector index 💬 Receives customer questions through a webhook 🔎 Searches Pinecone for the most relevant information 🤖 Uses OpenAI to generate a short, helpful answer 📤 Returns the chatbot response back to the connected frontend, chatbot, or app Key Benefits ✅ Turns a simple document into a searchable AI knowledge base ✅ Helps answer repeated customer questions faster ✅ Uses RAG so responses are based on the uploaded business information ✅ Works with any frontend, chatbot, or app that can send a POST request ✅ Easy to customize for different travel, booking, or support businesses ✅ Separates knowledge indexing from live chatbot responses ✅ Includes a simple response formatting step before returning the answer How It Works SW1 — Knowledge Base Indexing The workflow starts with a manual trigger. It downloads a knowledge document from Google Drive, loads the document content, splits it into smaller text chunks, creates embeddings with OpenAI, and stores those chunks inside a Pinecone vector index. This section should be run first because the chatbot can only answer from information that has already been indexed into Pinecone. SW2 — Chatbot Question Answering The webhook receives a customer question from a frontend, chatbot, or app. The workflow sends the question into a retrieval QA chain, searches the Pinecone vector store for relevant context, and uses OpenAI to generate a short answer based on the retrieved information. The response is then cleaned by a Code node and returned through Respond to Webhook. Features Google Drive knowledge document download OpenAI embeddings Pinecone vector storage Recursive text splitting Webhook-based chatbot input Retrieval QA chain for context-aware answers Short customer-support style responses Response cleanup before returning to the frontend Reusable structure for travel, booking, hotel, shortlet, and service businesses Requirements Google Drive account OpenAI API key Pinecone account and index n8n account or self-hosted n8n instance A prepared knowledge document such as FAQs, policies, service details, or booking information A frontend, chatbot, or app that can send questions to the webhook Setup Steps Connect your Google Drive, OpenAI, and Pinecone accounts. Add your own knowledge document to Google Drive. Create or select a Pinecone index for the vector database. Replace the Google Drive file placeholder with your own file. Replace the Pinecone index placeholder with your own index name. Run the indexing section once to store your document in Pinecone. Connect your frontend, chatbot, or app to the webhook URL. Send a test question using the chatInput field. Customize the system prompt for your business or client use case. Target Audience ✈️ Travel agencies answering booking and policy questions 🏨 Hotels and shortlet businesses handling repeated guest enquiries 🧳 Booking platforms that need a lightweight support chatbot 🤖 Automation builders creating RAG chatbot demos for clients 📞 Customer support teams that want answers based on internal documents. An n8n automation workflow template by theworkflowdev.
theworkflowdev
web
free
Others in the same category, ranked by how often they are opened.
