What this workflow does
This workflow answers customer questions about Yamaha’s 2024 Powered Loudspeakers brochure automatically via WhatsApp.
It saves time by replying fast and correctly without manual effort.
The system downloads the product brochure PDF, extracts the text, and turns it into a searchable format.
It listens for WhatsApp text messages, uses AI to find the right info, and sends answers back to users.
Who should use this workflow
This is for sales teams or businesses that get many repeated questions on WhatsApp.
It is useful for anyone who wants to save hours spent on manual replies and avoid mistakes.
The workflow fits Yamaha audio sales but can adapt to other product info in PDF brochures.
Tools and services used
- n8n Automation Platform: Runs the workflow and connects nodes.
- WhatsApp Business Account with n8n Nodes: Receives and sends WhatsApp messages.
- OpenAI API (GPT-4): Powers the AI assistant answering questions.
- HTTP Request node: Downloads the product brochure PDF file.
- Extract from File node: Gets text from the PDF.
- Recursive Character Text Splitter: Breaks large text into smaller parts.
- Default Data Loader: Prepares text data for embedding.
- Embeddings OpenAI node: Turns text into searchable vectors.
- In-memory Vector Store: Stores vectors for fast searching during chats.
- Switch node: Filters message types for proper handling.
How this workflow works: Inputs, Processing, and Output
Inputs
- HTTP request downloads the Yamaha 2024 loudspeaker brochure PDF.
- WhatsApp Trigger receives incoming WhatsApp messages.
Processing
- Extract text from the PDF file.
- Split text into small chunks to prepare for vector creation.
- Turn chunks into document format using Default Data Loader.
- Create embeddings with OpenAI and store them in the in-memory Vector Store.
- The Switch node routes text messages to the AI Sales Agent; other messages get an error reply.
- The AI Sales Agent uses GPT-4 along with vector store data and conversation memory to answer queries.
Output
- AI-generated answers sent back to users on WhatsApp instantly.
- Friendly message for unsupported message types.
Beginner step-by-step: How to build this in n8n
Import the workflow
- Download the workflow file using the Download button on this page.
- Open the n8n editor and select “Import from File.”
- Pick the downloaded workflow file to load it into n8n.
Configure the workflow
- Add OpenAI API Key in the AI Sales Agent and Embeddings OpenAI nodes.
- Connect WhatsApp Business credentials to the WhatsApp Trigger and reply nodes.
- Check the HTTP Request node for the brochure URL:
https://usa.yamaha.com/files/download/brochure/1/1474881/Yamaha-Powered-Loudspeakers-brochure-2024-en-web.pdf - Verify webhook IDs and phone numbers if needed for WhatsApp nodes.
Test and activate
- Run the workflow once inside n8n to test the PDF download and text extraction.
- Send a WhatsApp text message to trigger the AI response.
- Check that non-text messages get appropriate replies.
- Turn on the workflow toggle to activate it for daily use.
For self-host n8n setups, ensure the server and webhook URLs are public and accessible.
Customization ideas
- Change the brochure URL in the HTTP Request node to update product info.
- Adjust chunk sizes in the Recursive Character Text Splitter for better search results.
- Add support for images or voice in WhatsApp by extending the Switch node.
- Replace in-memory vector store with services like Pinecone or Qdrant for bigger scale.
- Edit system message in AI Sales Agent to fit your preferred tone or instructions.
Handling edge cases and errors
- If WhatsApp messages don’t arrive, check webhook URLs and WhatsApp API settings.
- If PDF text extraction is empty, verify PDF download and ensure Extract from File node is set to pdf extraction.
- If AI answers are wrong or too generic, rebuild the vector store and refine the AI prompt system message.
Summary and benefits
✓ Saves 3-4 hours of daily manual WhatsApp replies.
✓ Provides fast and fact-based answers using the brochure data.
✓ Reduces errors by removing human guesswork during chat.
✓ Boosts customer satisfaction with instant and accurate info.
→ Lets sales teams focus on closing deals instead of answering FAQs.
→ Workflow easily adapts by changing brochure or AI prompt.
