Two Cornell College students are working to enhance financial forecasting using artificial intelligence as part of the annual Cornell Summer Research Institute. Jillian Witt, a junior, and Liam Borer Seabloom, a senior, have teamed up with associate professor of finance Huan Cai. Their project aims to create a more efficient and user-friendly app that leverages large language models, such as ChatGPT, to parse and analyze retail sales data for companies seeking more accurate predictions.
Driven by a desire to improve industry standards, the students are not only exploring established forecasting methods used by leading firms but are also intent on making these advanced tools more accessible and affordable for a broader spectrum of users. Borer Seabloom, who studies finance and data science, is primarily responsible for the technical implementation, utilizing the OpenAI API and Python to process financial statements via the ChatGPT 4.1 model. Witt, a double major in finance and religion, is conducting competitive research and engaging directly with software developers to benchmark their new app against existing products and ensure it offers clear market differentiation.
The team’s faculty mentor, Huan Cai, highlights that this initiative targets the earliest and often most critical step in valuing companies—sales forecasting. By focusing on integrating large language models into this foundational process, the project aims to demonstrate that artificial intelligence can significantly affect the speed and precision with which financial analysis is conducted. With the Cornell Summer Research Institute running through July 11, the students hope to complete a prototype that offers meaningful improvements over traditional forecasting tools, with the goal of future public availability. Cornell College, located in Mount Vernon, Iowa, enrolls over 1,000 undergraduate students and serves as a launchpad for student innovation in applied research disciplines.
