AI-Driven Document Relevance Evaluation Workflow

This n8n workflow is designed to evaluate the relevance of documents retrieved from a vector store in response to specific questions. It integrates Google Sheets for dataset management, OpenAI for embeddings and relevance scoring, and LangChain components for processing and evaluation. The workflow initiates with a manual trigger, allowing the user to evaluate document relevance based on set metrics. When a question is received via a webhook, the system retrieves relevant documents from a Google Sheet, processes the documents into embeddings, and stores them in an in-memory vector store. It then compares the question against the stored documents, calculates a relevance score, and evaluates whether the retrieved information is pertinent. This setup is ideal for scenarios requiring automated and consistent evaluation of document relevance, such as chatbot training, knowledge base curation, or AI model validation.