Typing in concentrated haste, Megan Luby ’26 and her three groupmates eagerly race against the clock.
The Bryant undergrads upload an Excel spreadsheet of mock insurance sales data to Gemini, a generative artificial intelligence chatbot and virtual assistant, and let the tool analyze countless rows of information. As they wait for the AI to finish digesting the information, they consider how they will use the bot to compare this year’s sales to prior years based on office, region, and product segment and then derive important insights from that comparison.
“It's all about efficiency,” says Luby, noting that AI support will help them conduct analysis that would not be feasible given their allotted timeframe if done manually. But it’s up to the students to use the tool properly and to its best potential.
An Actuarial Mathematics major, Luby was one of 23 Bryant undergrads who visited Unum Insurance's Portland, Maine, headquarters in March to take part in the two-day Unum-Bryant AI Hackathon competition.
Working in groups, students used three different AI platforms — Copilot, Gemini, and ChatGPT — to gain experience with tools that will help define their futures and evaluate real data provided by the company.
“They're not just opening the door and letting us visit the company; they brought their practice to our students,” says Gao Niu, Ph.D., Mathematics and Economics department chair and professor, who helped organize the event and accompanied students alongside Professor of Mathematics and Economics James Bishop, Ph.D., and Assistant Professor of Economics Xiaofei Pan, Ph.D.
Flexing their AI muscle
Arriving at company headquarters on a Friday morning, undergrads talked with Unum employees about how the company is using AI in its day-to-day operations, the challenges they’ve faced, and the ways in which they aim to integrate AI into their workflows in the future.
“Getting to talk with their staff about some of the ways they've found AI beneficial and hearing their experiences was really cool,” Luby says.
The students were then introduced to their task. Bryant Actuarial Mathematics alum Colby Phillips ’23, who’s now part of Unum’s Actuarial Development Program and helped coordinate the Hackathon, notes that the data sets the undergrads used were based on real information.
“They're seeing what an actual data set and actual fields would look like,” says Phillips, noting that the only tweaks to the data sets were done to remove any sensitive identification information.
The Hackathon, Niu explains, gives students the chance to explore AI within the actuarial field but also to challenge the pervasive narrative that AI is “all knowing.” While the technology is evolving every day, humans are still needed to steer it and double check its conclusions, he says; a lesson that Luby’s group learned firsthand when they discovered that the way in which they prompted their AI chatbot led to errors with its calculations — prompting them to adjust their approach.
Drawing data-informed conclusions
Over the course of the Hackathon, each group — supported by an Unum representative — talked through sessions on initial analysis, planning, and exploration before creating final presentations.
“I had never used Gemini before, so experimenting with AI and knowing how to use the different tools to my advantage when I graduate and get a job will be beneficial,” says Luby.
In addition to analyzing datasets with AI, undergrads were also tasked with making recommendations to company leadership on how they could leverage AI tools to generate timely insights that would help them understand what to prioritize and why.
“Bryant focuses on presentation skills and working collaboratively, so it was good to see them flex that muscle that they have practiced using since their first semester,” shares Phillips.
Luby’s group — including Erik Tebaldi ’29, Zuzana Harvan ’28, and Jack Grady ’28 — won the best overall presentation distinction while a team made up of Sariah Hitchman ’29, Rain Lin ’27, Mason Day ’27, and Alexander Denson ’29 received an award for most valuable insight.
But all undergrads left with a newfound understanding of AI’s capabilities within the insurance industry.
“Applying quantitative skills with AI tools in a real-world setting like this is preparing them for the future of data-driven decision-making,” Niu says. “They have a new understanding of what the most powerful and popular AI tools are and how to use them to support your work and give you insight.”