Authors: Saransh Khanna, Mr. Ritesh Kumar Chandel
Abstract: This research investigates the development of an intelligent expense tracking system powered by Retrieval-Augmented Generation (RAG). Traditional expense tracking tools rely on structured inputs and static categorizations, creating friction for users who log expenses in natural language. The proposed system uses a hybrid architecture of vector retrieval and generative reasoning to extract accurate financial insights from unstructured text, improving reliability while reducing hallucinations commonly seen in standalone LLMs.
