Hi (or beep), I'm Lilo.
A chatbot assistant to support college enrollment
In this project, we designed a text-messaging chatbot that helps high school graduates navigate the college enrollment process. Our aim is to help reduce the phenomenon of “summer melt” where students who have been accepted to college decide not to attend at all. In our research, many reasons may underlie this decision to not attend college when already accepted. For some, key deadlines and paperwork is an obstacle. For others, social and emotional reasons that influence one’s college-going identity, play a role in not attending.
Lilo uses natural language processing to parse the various questions students may have before enrolling. For informational questions, we train Lilo on specific colleges’ information and provide just-in-time reminders and informational responses.
We’ve also designed Lilo’s persona and features to help students reflect on social and emotional concerns, and when needed, connect students to a “crowd” of available college mentors who can have quick text conversations about diverse issues that someone might face.
In 2022, we are partnering with NYU and CUNY to conduct an evaluation of Lilo with nearly 20,000 students. We’ll post results soon, as well as open Lilo for public use at other colleges.
If you want to know more about Lilo, we invite you to read our recent publication "A Mixed-Methods Analysis of Mechanisms to Support College Enrollment Among Low-Income High School Students". In this article, we build a better understanding of the issues that students encounter during the summer that deter them from enrolling in the first semester of college in the Fall.
We found that students needed not only information for completing college and financial aid applications but also an opportunity for social support and emotional guidance during the summer prior to college enrollment.