This n8n workflow automates the process of managing and querying company documents stored in Google Drive, integrating advanced AI services for semantic search and question-answering. The main goal is to streamline internal knowledge access by dynamically updating a vector store with new or modified documents and enabling users to ask questions related to these documents through a conversational AI interface.
The workflow begins with Google Drive triggers, which monitor a specific folder for newly created or updated files. When a change is detected, the file is downloaded, and its content is loaded into the system. The document content is then split into manageable chunks using a recursive text splitter, preserving context and avoiding information loss.
Next, the chunks are embedded into high-dimensional vectors using Google Gemini’s embedding model, and these vectors are inserted into a Pinecone vector database, which acts as an efficient semantic search index. For retrieval, the process involves querying the vector store with new questions, embedding these questions, and retrieving the most relevant document chunks.
A conversational agent, powered by Google Gemini’s chat model, interacts with users. It receives questions via a webhook, retrieves pertinent document information from the vector store, and provides accurate, context-aware answers in an engaging chat-based interface. The workflow also maintains context with a window buffer memory, enhancing the conversational experience.
Practical use cases include HR inquiries based on internal policies, quick access to technical documentation, or customer support scenarios where instant, accurate answers derived from company documents are essential. This workflow ensures that company knowledge remains dynamic, comprehensive, and easily accessible.
Overall, this automation saves time, reduces manual searching, and improves information accuracy within organizational processes.
Reviews
There are no reviews yet.