International Best Practices: How Are Universities Already Using AI-Based Student Administration Today?
An overview of international university success stories and proven practices. Daskalos Consulting, 2025–2026
Georgia State University's assistant, "Pounce," increased on-time enrollment rates by 3.3 percentage points and reduced the share of admitted students who never enrolled by 21%, according to a peer-reviewed, controlled study, at a cost of 7–15 dollars per student per year. This is just one, well-documented example that AI-based student administration is already delivering measurable results internationally today. Below, we present a few further best practices realized in university settings.
Georgia State University (USA): the most thoroughly studied case
Pounce originally launched in 2016 to address admissions and summer melt, meaning admitted students who fail to enroll. Its impact was measured by a peer-reviewed study (Page & Gehlbach, AERA Open, 2017): students with access to the system were more likely to enroll on time, and significantly fewer melted away between admission and actual enrollment. 85% of students responded to the assistant at least once, and only 13.5% of inquiries had to be escalated to a human staff member.
East Carolina University (USA): a replicated, confirmed result
A follow-up study conducted at another institution, also peer-reviewed (Page, Lee & Gehlbach, Research in Higher Education, 2021), further reinforced the picture: among first-generation students, loan take-up rose by 8 percentage points and enrollment by 3 percentage points. In an in-course experiment (Political Science 1101), students with chatbot access were 5–6 percentage points more likely to earn a B grade or higher, and 92% of students supported continuing the system.
Further international examples
California State University, Northridge (USA), "CSUNny": an admissions and onboarding assistant which, according to the institution's own tracking, was associated with higher retention and graduation rates among students with access.
University of Murcia (Spain), "Lola": an assistant answering admissions and enrollment questions, which, according to institutional data, handled more than 12,000 conversations and tens of thousands of inquiries within a single semester, with a high resolution rate.
Staffordshire University (UK), "Beacon": one of the first digital student coaching assistants in the United Kingdom, providing support on scheduling, administrative, and wellbeing questions to approximately 15,000 students.
Deakin University (Australia), "Genie": a comprehensive digital student assistant that handles up to 12,000 conversations per day at peak, and has won multiple international digital education awards.
Bolton College (UK), "Ada": a further-education institutional assistant that, according to institutional data, answered tens of thousands of student questions on scheduling, attendance, and library matters.
What do these best practices have in common?
A few patterns consistently recur across all these examples. The systems focus on concrete, well-defined processes, such as admissions, enrollment, or common administrative questions. They provide a clear, fast escalation path for cases they cannot answer independently. And they are available across multiple channels, continuously, not just on a single, isolated interface.
What does this mean for our own approach?
When developing Emily, we kept principles in mind similar to these international best practices: concrete, well-defined roles, source-cited answers, clear escalation, and full coverage of student and admissions processes within a single platform.
If you would like to see how these principles translate into practice in a demo, feel free to reach out to us.
Daskalos Consulting, 2025–2026. The international case studies presented in this article come from external sources, peer-reviewed studies as well as the institutions' own public communications. The full source list is available upon request.