Automated Grant Application Routing with RAG Agent

Overview

This workflow automates the process of routing grant applications. It begins with a Webhook Trigger to receive incoming application data. The data is then split into manageable chunks using the Text Splitter node and converted into embeddings using OpenAI's Embeddings node. These embeddings are stored and indexed in Supabase using the Supabase Insert node. When a new application arrives, the Supabase Query node retrieves relevant information based on the embeddings, which is then fed into a RAG Agent (Retrieval-Augmented Generation) alongside a Chat Model and Window Memory. The RAG Agent processes the application and relevant context to determine the appropriate routing. Finally, the outcome is logged to a Google Sheet using the Append Sheet node, and any errors are reported via Slack Alert.