Why AI matters at this scale

The University of Scranton is a private Jesuit university with over 135 years of history, serving approximately 3,800 undergraduate and 800 graduate students. As a mid-sized institution in the 1,001–5,000 employee band, it combines the agility of a smaller college with the complex operational needs of a comprehensive university. Its primary mission is delivering a values-based, liberal arts and professional education. In today's higher education landscape, institutions like Scranton face immense pressure: declining demographic trends, heightened competition for students, rising operational costs, and increased focus on measurable outcomes like retention, graduation rates, and career placement. For a university of this size, strategic technology adoption is no longer optional; it's a critical lever for sustainability and mission fulfillment. AI presents a unique opportunity to enhance the personal touch Scranton is known for, but at a scalable, data-informed level, allowing it to compete with larger research universities and more tech-savvy peers.

Concrete AI Opportunities with ROI Framing

1. Predictive Analytics for Student Retention: A significant portion of university revenue depends on tuition. Student attrition directly impacts financial health. An AI system that integrates data from the learning management system (LMS), student information system (SIS), and campus engagement platforms can identify students at risk of dropping out weeks or months before traditional methods. By alerting advisors to intervene with tailored support—academic tutoring, mental health resources, or financial aid guidance—the university can directly improve retention rates. A modest percentage point increase in retention translates to hundreds of thousands of dollars in preserved tuition revenue annually, offering a clear and rapid ROI on the AI investment.

2. AI-Enhanced Teaching and Learning: Faculty at mid-sized universities are stretched thin between teaching, research, and service. AI-powered tools can alleviate this burden and improve educational quality. Adaptive learning platforms can provide students with personalized practice problems and feedback, freeing instructors to focus on higher-order discussion and mentorship. AI teaching assistants can handle routine grading and FAQ responses for large introductory courses. The ROI here is twofold: it increases teaching efficiency (doing more with existing faculty resources) and improves learning outcomes, which boosts the university's academic reputation and attractiveness to prospective students.

3. Optimized Enrollment and Advancement: The admissions and development offices are revenue centers. AI can optimize prospect scoring for admissions by analyzing thousands of data points to identify applicants most likely to enroll, succeed, and align with the Jesuit mission, improving yield and academic fit. In alumni relations and fundraising, AI models can analyze giving history, career data, and engagement to predict donation propensity and personalize outreach. This moves advancement from broad, costly campaigns to targeted, efficient communications, increasing donor conversion rates and average gift size.

Deployment Risks Specific to This Size Band

For a university of Scranton's size, AI deployment carries distinct risks. Budget constraints are paramount; unlike massive research universities, Scranton cannot afford multi-million-dollar failures. Pilots must be small, focused, and demonstrate quick wins. Talent acquisition is another hurdle. Attracting and retaining data scientists and AI specialists is difficult and expensive, often requiring partnerships with external vendors or upskilling existing IT staff. Change management is particularly acute. With a strong culture and tradition, gaining buy-in from faculty and staff who may view AI as impersonal or threatening to jobs requires careful, mission-aligned communication and inclusive governance. Finally, data integration is a technical challenge. Legacy systems (SIS, LMS, finance) are often siloed, requiring significant middleware and cloud infrastructure investment before AI models can be effectively trained, creating upfront costs and complexity.