Answer product queries via WhatsApp using OpenAI GPT-4o and PDF knowledge base

WhatsApp AI Sales Agent using PDF Vector Store

This workflow turns your WhatsApp number into an intelligent AI-powered Sales Agent that answers product queries using real data extracted from a PDF brochure. It loads a product brochure via HTTP Request, converts it into embeddings using OpenAI, stores them in an in-memory vector store and allows the AI Agent to provide factual answers to users via WhatsApp. Non-text messages are filtered and only text queries are processed. This makes the workflow ideal for building a lightweight chatbot that understands your product documentation deeply.

Quick Start: 5-Step Fast Implementation

1. Insert your WhatsApp credentials in the WhatsApp Trigger and WhatsApp Send nodes. 2. Add your OpenAI API Key to all OpenAI-powered nodes. 3. Replace the PDF URL in the HTTP Request node with your own brochure. 4. Run the Manual Trigger once to build the vector store. 5. Activate the workflow and start chatting from WhatsApp.

What It Does

This workflow converts a product brochure (PDF) into a searchable knowledgebase using LangChain vector embeddings. Incoming WhatsApp messages are processed and if the message is text, the AI Sales Agent uses OpenAI + the vector store to produce accurate, brochure-based answers.

The AI responds naturally to customer queries, supports conversation memory across the session and retrieves information directly from the brochure when needed. Non-text messages are filtered out to maintain clean conversational flow.

The workflow is fully modular: you can replace the PDF, modify AI prompts, plug into CRM systems or extend it into a broader sales automation pipeline.