Of the 2.62 million students who started college in fall 2024, 86% came back for a second semester and 77% were still enrolled somewhere the following fall, according to the National Student Clearinghouse Research Center's June 2026 persistence and retention report. Roughly one in four first-year students left higher education entirely within 12 months.
A lot of research has been written about why students leave higher ed and what colleges and universities can do to improve retention; often these efforts focus on the biggest issues students face that lead them to drop out, namely financial stress and difficulty with both the transitions and expectations of college. And while technology is not the answer to every ailment, it can (and does) play a significant part in supporting students in meaningful ways across higher education.
There are five main ways that universities use technology to improve learner outcomes: they flag students who are drifting before those students disappear, they build participation into large lecture courses, they offer enrollment options that fit around jobs and caregiving, they design courses so the material is usable for every student who enrolls, and they rebuild assignments around what AI has changed about assessment.
Early alerts and advising to support students
A lot of promises about what technology can do are more aspirational than practical, but one thing technology is notably good at is recognizing patterns and reporting on data. That means you can use it to watch student records for the specific patterns that preceded a withdrawal in past cohorts, then take action.
Georgia State University launched its GPS Advising system in August 2012 and runs more than 800 checks on undergraduate records daily, looking for things like a missing prerequisite, a low grade in a major-critical course, or a registration that doesn't move a student toward a degree. When a check trips, an advisor reaches out. The university reports more than 55,000 one-on-one advising meetings in a single academic year, and freshman fall-to-spring retention up five percentage points since GPS launched. In 2018, Timothy Renick, then Georgia State's senior vice president for student success, reported that graduation rates had risen 23 percentage points since 2003 and that gaps by race, ethnicity, and income had closed.
Having the data doesn’t drive outcomes by itself; the follow-up is where the outcomes come from. Georgia State staffed advising to match the volume of alerts its system produces, which is key. The technology is a tool, but the school builds the systems that make it a useful one.
Participation built into large courses
Lecture halls are big places (the word hall really distinguishes them from, say, a high school history classroom). They’re designed for volume, to hold a lot of students and deliver a lot of information at the time when education goes from schools of several thousand to universities of tens of thousands. But it can be a challenging format for both instructors and students; an instructor teaching 400 students might not be able to tell who's following along until the midterm arrives, where students feel like just a number in a sea of faces.
Students who feel that disconnect are more likely to consider leaving. Gallup research from 2024 reported that 35% of students considered leaving their program, with emotional stress, mental health reasons, and feelings of not belonging among the top five reasons behind considering a pause.
Indiana University started working on the lecture hall engagement problem in 2017 by connecting Top Hat, a student engagement platform, to Canvas. The integration runs on LTI 1.3, so students open Top Hat from inside their Canvas course and grades, assignments, and content sync back without anyone rekeying them.
Instructors embed questions inside assigned readings, run polls and quizzes during the session, and set low-stakes assessments that register participation in the moment. In Basic Human Anatomy, Dr. Valerie O'Loughlin uses polling, quizzes, and interactive prompts. In Pulmonary Grand Rounds, Dr. Graham Carlos runs case-based learning through interactive scenarios.
The program now spans nearly 1,900 courses and more than 51,000 students. Students in Top Hat-enabled courses withdrew at a rate 11.5% lower than their peers, which Instructure's account of the rollout puts at roughly 290 students who stayed on a path toward a degree. In O'Loughlin's anatomy course, students who used the tool averaged 79.4% on exams against 46.7% for students who didn't.
Meanwhile, one Canadian university points directly to the learning management platform as driving engagement among students, with each other and with the learning itself. Martin Laba, Associate Professor in the School of Communication at Simon Fraser University, said:
“The engagement and freedom students feel is palpable. Students are no longer passive recipients of knowledge. They want to co-create content, deeply and actively engage with course material, and even at times revise course content and direction… We’ve opened up a lively public sphere in Canvas, a space of discussion and debate for students. It’s a space that gives depth and substance not only to the content but to the quality of the student experience in the course. They can send the course in new directions each week. They can urge classmates and colleagues to not just analyze but engage through real-world actions of consequence.”
Enrollment options that fit around work (and the rest of life)
In Gallup’s 2024 research, cost came up as another top consideration for students thinking about leaving their programs. For those students trying to balance higher ed with real financial pressure, the question they’re trying to answer is, “can I take this course without quitting my job?” Others are juggling caregiving, or they got into their dream school and can only afford it by commuting two hours each way.
In short, a student working 30 hours a week and a student caring for a parent both need asynchronous access to lectures, and the technology that provides it removes a reason to stop out.
Fall 2024 federal IPEDS data shows that of 19.6 million students enrolled in U.S. higher education, 26.5% took courses exclusively at a distance and another 28.2% took at least one. An LMS that supports remote learning does a specific set of things:
- Asynchronous course delivery. Content, assignments, and quizzes sit in one place for students on their own schedule, which lets someone stay enrolled through a job, a caregiving shift, or a time zone.
- Video that lives inside the course. Lectures and demonstrations stay attached to the material they belong to, so a student who missed the live session can pick the week back up.
- Grading and feedback in the platform. Marked work, annotations, and comments reach students without paper copies changing hands.
- Activity and progress data. Instructors and advisors can see who's gone quiet, which is most of the visibility they have when nobody's in the room.
- Mobile access. Students working from a phone stay current.
- Integrations with grade passback. Publisher content, proctoring, engagement tools, and accessibility checkers run inside the course, so students have one login and one place to go.
- Shared course structure. Every section works the same way, so students spend their attention on the material.
Access can also be a catalog problem. The Montana University System serves 32,000 students across 15 campuses and seven tribal colleges, spread across a state that takes 10 to 12 hours to drive. Its smaller campuses can't always offer every required course. In 2025, they put Parchment Course Sharing inside Canvas, so a student can enroll in a course from another campus and still get one bill from their home institution. The system's own analysis found that for three-quarters of its institutions, sharing online catalogs would more than double the unique lower-division courses available in a typical term.
Course design that works for more learners
When 23% of students report feeling like they don’t belong and 7% say they felt discriminated against or harassed, accessibility in edtech becomes about more than just compliance. While the Department of Justice's ADA Title II rule sets WCAG 2.1 Level AA as the technical standard for public institutions' web content and mobile apps, accessibility goes beyond legal accommodations and simply offers more support to more students. Captions, real heading structure, readable contrast, and documents that work with a screen reader are useful to students, whether or not that student has ever filed for an accommodation, because the tools that make a course usable also make it easier to engage with.
The Canvas Accessibility Checker flags problems while an instructor is still building a page, and gives students a high-contrast interface and Immersive Reader so they can change how content reaches them. At Mt. San Antonio College, educational technologist Eric Turner says text-to-speech through ReadSpeaker in Canvas works because it's simple: one Listen button, with students choosing voice and speed.
AI that has moved into assignment design
Institutions are facing two truths about AI in education at the same time: A student who lets AI do the thinking hasn't learned what the degree is meant to teach, and a student who graduates without knowing how to work with AI walks into a job market that's already pricing the skill.
From a preprint from the MIT Media Lab, in a study led by Nataliya Kosmyna, participants wrote essays under three conditions (using an LLM, using a search engine, or using nothing) while EEG measured brain activity. Participants working without tools showed the strongest and most distributed neural connectivity, and LLM users showed the weakest. LLM users also reported the lowest sense of ownership over their essays and struggled to quote work they had just written. The authors call this “cognitive debt.”
The second concern is really about the job market. Lightcast analyzed more than 1.3 billion job postings for its July 2025 report and found that postings requesting AI skills carry a 28% salary premium, close to $18,000 a year, and that as of 2024, 51% of postings requiring AI skills sit outside IT and computer science.
So students need to be able to learn about AI, as well as learn without AI. One higher ed focused study showed that roughly one-quarter of faculty are now redesigning assessments AI can't answer for them, focusing on methods like oral exams or interactive, simulation-based assignments.
James Genone, a philosopher and cognitive scientist who led learning strategy at Northeastern University, believes that institutions will fail (themselves and their students) at AI when they give faculty no guidance and when they hand down a mandate outright banning it. What works is faculty building shared frameworks together. In his work at Northeastern, he brought together groups of faculty and let early adopters teach their peers as voices their colleagues already trusted. He set up working groups to build shared frameworks for questions like AI readiness and assessment. Department and college leaders then carried those frameworks into courses, so students got a consistent experience across programs, allowing every student to engage with it in a way that will build their AI literacy.
The why behind what’s working
Technology doesn't automatically move outcomes on its own; someone is sitting behind the software. Georgia State's 800 daily checks matter because advisors answer them. Indiana University's participation data matters because instructors change what they do the next week. Genone's frameworks matter because faculty built them together. Software's job is to surface a problem early enough that someone can still do something about it, and everything after that is people, and the decisions institutions make.
Early alerts, participation data, flexible enrollment, accessible design, and rebuilt assignments are five responses to a much longer list of pressures on higher education right now. Find out more about how technology can be used to think about approaching challenges in higher ed with our The Future of Edtech: Building Education For The Lifelong Learner report and our The Future of Edtech: Building Better AI for Education report.
Frequently asked questions
How does technology improve student outcomes? Technology improves student outcomes by surfacing problems while there's still time to act on them and by removing barriers to finishing. Early alert systems flag students who are off track, participation tools show instructors who's disengaging in large courses, online options let students stay enrolled through work and family obligations, and accessible design keeps course materials usable. In every case, the outcome depends on a person responding to what the technology shows.
What technology do universities use to support students? Most institutions run a learning management system as the hub, with early alert and advising platforms, learning analytics dashboards, accessibility checking tools, and student engagement or polling tools connected to it. The specific mix varies, though the LMS is usually the system of record because it's where course activity already happens.
Do early alert systems actually improve retention? They can, when institutions staff the response. Georgia State University reports freshman fall-to-spring retention up five percentage points since launching its GPS Advising system in 2012, alongside more than 55,000 one-on-one advising meetings a year. Results at other institutions vary widely, and the variation tends to track advising capacity more closely than software choice.
