Two Cornell Summer Research Institute (CSRI) students are diving into the AI world of machine learning to create a more effective tool for financial analysts to forecast sales.
Rising junior Jillian Witt ’27 and rising senior Liam Borer Seabloom ’26 are working with Associate Professor of Finance Huan Cai to incorporate large language models (LLMs)—such as ChatGPT—in forecasting sales for retail companies. Their goal is to create an app that’s easier for folks in the financial field to use.
But with the pace of change, the research team knows they have a lot of work to do.
“AI can drive innovation, solve complex problems, enhance automation, and productivity,” says Witt, who is a finance and religion double major. “To understand the benefits of AI and its potential drawbacks, we need to conduct this research.”
Sales forecasting, or predicting how much money a company expects to generate in the future, is nothing new. But incorporating LLMs into that process is.
“Right now, there are a ton of apps that are being used and are sold to the public, that are doing sales predictions,” says Borer Seabloom, a finance and data science double major. “We want to build an app that incorporates AI, as well as the methods that these other companies are using, and see if we can improve upon their current methods and make them more publicly available or cheaper.”
To do this, Borer Seabloom is putting his coding skills to use, and even though it’s early in the research process, he’s finding success.
“So far, I’ve been working in Python with OpenAI’s application programming interface (API),” Borer Seabloom says. “We’ve been prompting the ChatGPT 4.1 large language model, which is a pretty recent model, and we’ve been able to get out easy-to-read and well-formatted information from a financial statement that it’s been given.”
Witt, on the other hand, is practicing her personnel and management skills through requests for demos from other software developers who have similar products. The CSRI team wants to ensure their creation is more advanced and not a duplicate of another product. Witt says this eight-week research process will support her dreams of becoming a business owner and funeral director.
“I really like working with Huan,” Witt says. “I appreciate how she’s really looking out for me, not just now, but in the future. This opportunity will help me further develop critical thinking skills and allow for career exploration. It is also important to understand AI as it becomes increasingly relevant in our day-to-day lives, particularly in my field of interest.”
The finance professor points out that this project is part of a much bigger process that the students will need to learn and understand as they head out into the workforce.
“Sales forecasting is actually the first, substantial step of any equity valuation,” Cai said. “We need to know the sales of the companies, then we forecast expenses, and get the profit of the company. With the profit, we calculate the company’s future cash flow, and eventually, we would get the theoretical value of a stock. We are doing the very first step to see how LLMs can play a role because I think that’s the future that professionals will move to once we have the large language model combined with the finance practice.”
Cai notes that employers value research experiences on a resume and often require candidates to be able to read and understand academic papers. She also emphasizes that LLMs can be very useful in some settings but problematic in others, so it is important to find out when, where, and how to incorporate LLMs in the sales forecasting process.
“It’s been great to work on this project because it is a combination of my two majors, data science and finance,” says Borer Seabloom. “I think CSRI is just a wonderful way to get experience in both of those fields, which I want to work in eventually. My dream job would be working as a lead data scientist at some company in the Twin Cities, doing data analysis and modeling for them.”
This is just one of many research projects taking place on the Cornell College campus this summer. CSRI runs from May 19–July 11.
Click below to watch the video on this CSRI project:
