Turn WhatsApp into a smart sales chat channel. Customers send questions and an AI agent replies with clear answers pulled from your product brochure. Great for sales teams that need fast, accurate replies without a live rep on every chat.

Behind the scenes, a manual path loads your brochure once. It downloads a PDF, extracts text, splits the content, creates embeddings with OpenAI, and stores them in an in memory vector store. The live path starts when a WhatsApp message arrives. Non text messages get a friendly notice. Text questions go to an AI agent powered by an OpenAI chat model, a memory node that tracks each phone number, and a vector store tool that searches the brochure for the best facts. The reply is sent back to the customer on WhatsApp.

You will need a WhatsApp Cloud API setup and an OpenAI API key. Run the manual step to build the catalog, then activate the workflow. Expect reply times to drop to seconds and support for many chats at once. Use it for product Q and A, spec checks, and guided buying help using the exact wording from your brochure.

• WhatsApp Trigger listens for incoming customer messages and starts the chat flow.
• Message type handling filters non text content and sends a polite notice.
• HTTP Request downloads the latest product brochure PDF from a URL.
• Extract from File reads PDF text for use in search and answers.
• Recursive Text Splitter breaks long text into chunks for better search results.
• OpenAI Embeddings converts each chunk into vectors for similarity search.
• In memory Vector Store holds your product catalog for quick lookups.
• AI Agent uses an OpenAI chat model, session memory per phone number, and a vector tool to answer questions.
• WhatsApp Send node returns the agent’s response to the customer as a text message.
• Manual Trigger lets you run the catalog build step once and update it when the brochure changes.

• Reduce manual replies from minutes to under 10 seconds
• Handle up to 10 times more concurrent chats with the same team size
• Improve answer accuracy by citing facts from your brochure
• Eliminate most copy paste work from sales reps
• Keep context per customer so follow up questions stay on track

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with WhatsApp and OpenAI. See the Tools Required section above for links to create accounts with these services.
3. Open the WhatsApp Trigger node in n8n. In the Credential to connect with field, click Create new credential and follow the on screen instructions to connect your WhatsApp account and verify the webhook in your Meta app.
4. In the WhatsApp Trigger, make sure updates is set to messages so incoming chats are captured.
5. Open the WhatsApp Send nodes and replace the phoneNumberId with your own number ID from the WhatsApp Cloud API.
6. Open the OpenAI nodes. In the credentials dropdown, click Create new credential. Paste your OpenAI API key from your OpenAI account and save.
7. Optionally change the brochure URL in the HTTP Request node if you want to use your own PDF. Keep Extract from File set to pdf.
8. Click the Manual Trigger and run the left side of the workflow to download the PDF, extract text, and build the in memory vector store. Look for a successful run without errors.
9. Send a test message to your WhatsApp number. A text message should receive an AI answer. A photo or audio should get the non text notice.
10. Open the AI Sales Agent node and confirm the memory key uses the sender phone number so the chat stays personalized.
11. Activate the workflow in n8n so the WhatsApp Trigger can receive messages when n8n is not in test mode.
12. Troubleshooting: If answers seem empty after a restart, rebuild the catalog because the vector store is in memory. If the trigger does not fire, recheck the webhook subscription in your Meta app and ensure your server URL is reachable.