Spend Sense
Description Built an AI-powered expense intelligence system that extracts and categorizes financial transactions from unstructured notification data using LLMs. Enabled natural language querying over personal finance data through an agent-based workflow for actionable insights and tracking.
Problem Statement - Users receive financial information (SMS, notifications, emails) in unstructured formats, making it difficult to track, categorize, and analyze expenses effectively. This leads to poor financial visibility and decision-making. Additionally, users lack intuitive ways to query and understand their spending patterns without manual effort.
Solution - Develop an AI-driven system that ingests unstructured financial data, uses LLMs to extract and classify transactions, and stores them in a structured format. Implement an agent-based interface that allows users to query expenses in natural language, generate insights (spending trends, category breakdowns), and improve financial awareness with minimal manual input.
