AI-Powered Financial Wellness Coach

Overview

The AI-Powered Financial Wellness Coach is a prototype web application designed to provide personalized financial guidance and support to users of the Michigan State University Federal Credit Union (MSUFCU). The application utilizes Large Language Model (LLM) technology to analyze user input through a chat interface and generate tailored financial advice, helping users make informed decisions about their finances.

A primary feature of the project is its integration with a Model Context Protocol (MCP) server, which allowed the application to query financial information using specified processes through a secure channel. This allowed the coach to be more personalized to a specific user's financial situation, as well as allowed it to reference and guide users to a verified source of information at MSUFCU's financial information platform, all while ensuring the proper security and privacy of the user's data.

The project included a web interface using React to allow users to interact with the AI coach, as well as a backend server to handle requests and manage the conversation context. The web interface was designed to be user-friendly and accessible, providing a seamless experience for users seeking financial advice.

This project was developed as a prototype for MSUFCU as part of a team of six people for our capstone project, on behalf of Michigan State University. My primary contribution to the project was most of the workflows related to connecting to and conversing with the AI using Python, including retaining memory and properly calling functions and MCP tools. I also set up most connections between the different services, as well as assisted with the proper display of the web interface.

You can find an archive of the project information on the capstone website here.