Eight nonprofit colleges have announced closures in 2026, on top of sixteen that closed in 2025. Hampshire College, Anna Maria College, Sterling College, Lourdes University, and Labouré College of Healthcare are among them, each citing some version of the same combination: declining enrollment, rising costs, and reserves that finally ran dry.
They will not be the last.
New research from Huron Consulting Group projects that 442 of the nation's roughly 1,700 private nonprofit colleges and universities, representing about 670,000 students, are at risk of closing or merging within the next decade. More than 120 of those institutions sit at the very highest risk level. The number of 18-year-old high school graduates peaked at about 3.9 million in 2025 and is projected to decline annually for the next fifteen years. New international student visa issuances fell roughly 36 percent for the 2025 to 2026 intake. Nearly a third of private nonprofit colleges posted operating deficits in 2024.
This is not a warning about the future. It is a description of the present.
And yet most institutions facing these pressures are still operating with the same manual processes, fragmented data systems, and reactive decision-making models they have relied on for decades. The schools that survive the next ten years will not be the ones that simply cut deeper. They will be the ones who change how they use information to decide, allocate, and serve students. The smart deployments are not about headcount. They are about capability, and about protecting the revenue you already have.
That is where integrating AI becomes a financial strategy, not just a technology initiative.
The winning approach is already settled, and it is human-led
Before the operational case, it is worth naming the consensus, because it is not the consensus most vendors are selling.
EDUCAUSE, the higher education technology association whose research your IT and academic leadership already reads, ranked "The Human Edge of AI" as the number-two issue in its 2026 Top 10. Its guidance on which institutions will actually thrive is direct: the winners are those with adaptive cultures, strong leadership trust, and a commitment to using technology to enhance, not replace, the human elements of education.
That is the entire thesis of how we work. The technology is not the product. Adoption is. A university does not have an AI problem. It has an adoption problem, layered on top of faculty governance, accreditation requirements, and the real ways decisions get made on a campus. Our model is Human Led, AI Facilitated, and the evidence below shows what that looks like when it is done right and what it costs when it is not.
Retention is a revenue problem, and the math is public
Start here because this is the clearest dollar case in all of higher education, and it has been audited publicly for more than a decade.
Georgia State University began using predictive analytics for advising in 2012. The system tracks every undergraduate daily against more than 800 risk indicators, and when a student earns a low grade or shows another at-risk signal, an adviser is alerted to intervene early rather than after the damage is done. The results are not vendor estimates. They are Georgia State's own reporting.
By the university's calculation, every 1 percentage-point increase in retention is worth about $3.18 million a year in additional revenue. Tim Renick, who led the work and now directs Georgia State's National Institute for Student Success, has put the graduation-rate figure even higher: every percentage point of improvement is worth roughly $3 million a year in tuition and fees, and since 2003 the university has raised its graduation rates by 23 percentage points, which adds up to more than $60 million.
Be precise about what actually drove those numbers, because the honest version is the stronger one. Georgia State did not simply buy software and watch the revenue arrive. It also added advisers and ran a completion-grant program for students at financial risk. The analytics were not a substitute for those people. They were what made those people effective. With one adviser for every thousand students, the staff could never have found the right students at the right moment on their own. The system pointed scarce human attention at the students who could be helped, exactly when intervention still mattered. That is not a labor-replacement story. It is the human-led model proven: the data tells the adviser where to look, and the human does the advising.
That distinction is also why this matters more for a small college, not less. A 1,200-student institution cannot hire 42 advisers. But it does not need to. It needs a system that lets the six advisers it already employs reach the students who can actually be saved, rather than spreading attention thin across everyone or, as is usually the case, only catching students once they are already failing. And at that scale, every retained student sits closer to the line between solvency and closure, so the leverage on each one is higher. The platform Georgia State used costs on the order of $150,000 a year against tens of millions in recovered revenue. The tooling is not the expensive part. Misdirected human effort is.
Enrollment: reach the students you already admitted
The average institution spends thousands of dollars per enrolled student on recruitment and marketing, much of it untargeted. But a large share of lost enrollment is not a recruiting failure at all. It is summer melt, the students who are admitted intend to come, and then never make it to the first day because a financial aid form, an immunization record, or a registration step quietly defeats them.
Georgia State again offers the proof point. Its AI assistant, named Pounce, communicates with admitted students by text, answering questions around the clock and nudging them through each enrollment task. The university reduced summer melt from 19 percent to 9 percent. The efficiency story is captured in a single line from Scott Burke, the university's assistant vice president of undergraduate admissions: "We would have had to hire 10 full-time staff members to handle that volume of messaging without Pounce." And it did not bury the staff it had. Of more than 50,000 student messages received, fewer than 1 percent required a human staff member, which freed counselors for the cases that genuinely needed professional judgment.
This is the responsible version of enrollment work, and it is worth being precise about what it is and what it is not. The goal is fit and follow-through, helping students who choose your institution actually arrive and persist, and packaging aid in a way that supports yield without over-discounting. It is not about screening students out or chasing the ones most likely to pay. Syracuse University's recent experience, discussed below, is a reminder of how badly the extractive version of that strategy can backfire.
Program portfolio decisions: proactive beats reactive
Most institutions cut programs reactively, after years of enrollment decline have already drained the resources, at which point the cuts feel like crisis management. There is a better version, and Syracuse University recently demonstrated it.
In April 2026, Syracuse announced it would sunset 93 academic programs as part of a deliberate Academic Portfolio Review. The provost made it clear that this was not a cost-cutting exercise driven by financial necessity. Of the 93 programs, 55 had zero students enrolled, and the cuts affected only about 1.2 percent of the student body, all of whom can finish their degrees. Syracuse was carrying roughly 460 degree and certificate programs, well above the peer average of about 200. That is what proactive portfolio intelligence looks like: deciding where to invest, sunset, or redesign based on real demand and contribution, before any single program becomes a financial liability.
The cautionary half of the Syracuse story came separately, two months later. In June 2026, the university disclosed its first budget deficit in years after missing its fall enrollment target, a shortfall tied in part to a financial aid strategy that lowballed committed families while extending large packages to students who had already declined. That is the failure mode AI-assisted aid modeling exists to prevent. The institution could have modeled the yield and revenue impact of those aid decisions before making them.
The lesson for leadership is the distinction itself. The portfolio review was strategic and defensible. The aid-driven deficit was reactive and avoidable. Integrating AI is what moves more of your decisions into the first category.
Note: This article was researched and written by Justice Jones with AI assistance, then reviewed and edited by our team. External studies and sources are credited to their original authors. Examples from our own work reflect our organizational practice.
Administrative efficiency without gutting your workforce
Small colleges run the same back-office processes as large research universities, with a fraction of the staff. Financial aid verification, transcript evaluation, scheduling, compliance reporting, and accreditation documentation all consume enormous human time, and errors carry real financial and regulatory consequences.
This is no longer a fringe idea. EDUCAUSE ranked administrative simplification as the number-two issue on its 2025 Top 10, and institution-wide AI adoption across higher education jumped from 49 percent in 2024 to 66 percent in 2025, with 88 percent of institutions expecting use to continue rising. Industry experience suggests that AI-assisted automation can meaningfully reduce overhead in targeted areas such as document processing, aid verification, and scheduling optimization. We present that as an illustrative range rather than a guarantee, because the honest state of the field, as EDUCAUSE itself notes, is that ROI measurement for these initiatives is still maturing. What we will commit to is a scoped pilot with a defined baseline, so the savings are measured rather than assumed.
The point is not doing more with less. It is to stop wasting skilled human labor on tasks that technology handles more accurately and faster, so that staff can return to direct student support and strategic work.
Prove the value to the student, not just to the institution
Public confidence in the payoff of a degree has fallen sharply over the past decade, and families now ask a blunt question: Will this investment lead to a career?
AI can help an institution answer that with specificity, and this is the constructive inverse of yield-chasing. By matching an incoming student's profile to the outcomes of alumni who started in similar circumstances, a school can show a prospective student a credible, personalized path. A first-generation student from a rural community considering nursing can see the trajectories of alumni who entered with comparable backgrounds and went on to licensure and employment. This is not a generic 95 percent placement claim. It is evidence aimed at the student's decision, demonstrating value rather than screening the student out, and it directly answers the skepticism that now shadows the enrollment conversation.
How we work, and why it lands
We pair integrating AI with structured change management, because the technology only delivers if the people and processes around it are ready to adopt it. This is the work 24/7 Teach has done with more than 50 organizations: not arriving with a platform and a promise, but designing phased adoption that aligns with governance, accreditation, and how decisions actually get made inside an institution. In practice that means starting with a readiness assessment. We interview your leadership, from the president's goals down to the operational reality your provost, CFO, CIO, and enrollment lead actually live with, and turn it into a prioritized roadmap with one pilot scoped and ready to run. From there you expand from evidence rather than enthusiasm. Universities do not buy moonshots. They buy a scoped first step that de-risks the next one.
That sequencing is also what protects the thing that matters most. Used well, as Georgia State's leaders describe their own work, this is high tech and high touch. The technology earns its place by making the people you already have more precise, not by standing in for them.
What this does not do, and what to do next
Two honest caveats, because you have heard salvation pitched before. First, AI does not create demand. It will not manufacture more 18-year-olds or restore public faith in a degree. What it does is protect the revenue from the students you have already enrolled and gives you a runway to address the demand problem with everything else in your strategy. Second, it is one lever among several. A merger, a portfolio bet, or an online expansion may be a higher-return move for your institution than anything in this article, and we will tell you plainly when that is the case rather than sell you a pilot you do not need.
What integrating AI does offer is the difference between making decisions with full visibility and making them in the dark. The institutions that act now, while they still have the resources and runway to do it well, hold a structural advantage over those that wait until the options have already narrowed. The schools that closed in 2026 did not fail because their leaders were not working hard enough. They failed because the math stopped working, and by the time it was visible, it was too late to change it.
The question is not whether your institution will need this. It is whether you implement it proactively, on your terms, or reactively, on someone else's.
If you want to see what this looks like on your campus, the place to start is an AI Readiness and Roadmap engagement. We interview your leadership, map where the financial math is most exposed against the goals you are accountable for, and hand you a prioritized roadmap with one pilot scoped and ready to run. You get a decision-ready plan you can take to your board, not a sales deck. Tell us your three priorities for the year, and we will show you where AI moves them and where it does not.
About the author
Justice Jones is an instructional designer, AI strategist, and former K-12 principal, and the co-founder and CSO of 24/7 Teach. He built the company to close the gap between what schools teach and what teens and professionals need to succeed, and he leads AI strategy at its sister company, Naomi-AI, a K-8 classroom platform. Through 24/7 Teach, he and his team have supported more than 50 organizations and placed more than 600 adults in new careers.