This n8n workflow creates an automated customer support system that leverages AI to fetch and answer customer queries via WhatsApp. The system continuously crawls a company’s website to retrieve up-to-date information, ensuring accurate and real-time responses. The workflow begins with a WhatsApp trigger to detect incoming messages. It then uses an AI agent, powered by a language model, to analyze the message content and determine the best way to find the answer.

The AI agent utilizes specialized tools to crawl the company’s website, starting from the root domain and extracting internal links. It selectively retrieves relevant pages based on keywords related to common customer inquiries like product details, shipping, or policies. The retrieved page content is processed and cleaned for better readability.

To manage ongoing conversations, the workflow checks if the interaction occurred within the last 24 hours. If so, it responds directly with the extracted answer; otherwise, it can send a pre-approved template to reopen the communication. Responses are sent back to the user through WhatsApp, providing timely and accurate support.

This setup is especially useful for eCommerce or service websites that require dynamic, real-time customer queries handling without manual intervention, saving time and enhancing customer satisfaction.

Node Count	11 – 20 Nodes
Nodes Used	@n8n/n8n-nodes-langchain.agent, @n8n/n8n-nodes-langchain.lmChatOpenAi, @n8n/n8n-nodes-langchain.memoryPostgresChat, @n8n/n8n-nodes-langchain.toolHttpRequest, code, if, stickyNote, whatsApp, whatsAppTrigger