This n8n workflow creates a sophisticated, multi-agent chatbot system tailored for different websites and pages. When a user sends a query through a webhook, the workflow dynamically determines the site and page context using sticky notes and a switch node. It then routes the query to specialized AI agents that utilize LangChain’s language models, Pinecone vector stores, and SQL chat memory to generate accurate, context-aware responses. Each agent is configured with a custom system prompt tailored to the specific site or page, ensuring the reply aligns with the content and services offered.
The workflow includes multiple embedded AI agents, each connected to different Pinecone indexes representing various data domains like services, terms of use, privacy policies, and refund policies. These agents leverage vector searches to retrieve relevant information, which the language models utilize to formulate precise answers. Responses are then sent back through webhook response nodes.
This system is practical for multi-site, multi-page customer support, providing users with instant, accurate answers based on up-to-date data stored in Pinecone vector databases. It ensures scalable, context-specific interactions for websites with diverse content and service offerings.
Reviews
There are no reviews yet.