Every institution is losing revenue somewhere. Leads never called back. Students who dropped in year two. Financial aid families who walked away because no one explained the next step. Six-figure pre-collections write-offs treated as a rounding error. The question isn’t whether the leaks exist. It’s whether AI finds them in time for the institution to act.
In Episode 4 of AI Talks, Sathish from Binary Words sits down with Aaron Vasko, Chief Multiplier at Mac Insight Group, to unpack what AI means for enrollment, financial aid, retention, international recruitment, and the broken financial model underneath roughly 80% of US colleges. The phrase Aaron keeps returning to: your boat is leaking. Not as a warning. As a diagnosis.
Aaron Basko
Aaron Basko is Chief Multiplier at Mac Insight Group, a higher-education consulting firm. With 30 years in enrollment, including two vice presidencies and consulting work for over 20 universities, he helps institutions strengthen enrollment, financial aid, retention, and international student recruitment strategy.
Sathish Kumar Mariappan
Sathish Kumar Mariappan is Co-Founder of BinaryWorks, an Atlanta-based certified Drupal agency. He helps education, government, healthcare, and enterprise organizations with Drupal development, migration, AI consulting, and digital strategy, building secure, scalable digital experiences across complex ecosystems.
Episode TL;DR
- → AI is moving from chatbot to strategic advisor. It informs leaders on what to share at board meetings and which students to call today.
- → Students search with ChatGPT and Claude, not Google. Institutions invisible in AI-native discovery are losing top-of-funnel reach.
- → Speed to lead is measured in minutes, not days.
- → Financial aid is the slowest adopter but holds the biggest gains in pre-admission packaging, self-service, and pre-collections outreach.
- → Retention is stuck near 50% nationally. A 2 to 3% lift outperforms most marketing spend on ROI.
- → Smaller institutions can use AI as an equalizer if they build to their specific funnel.
- → International recruitment (Latin America, online dual enrollment) is the most under-exploited opportunity.
- → The old revenue model is broken. “More students plus small tuition increases” no longer fits the environment.
- → One Mac Insight client recovered $5.6 million in pre-collections in two years by reaching out before bills hit collections.
From Chatbots to Strategic Advisors
Sathish: How has AI changed the landscape of enrollment?
Aaron: “We started with chatbots, and they’ve gotten more intelligent. Instead of just texting, you could get them to phone call a student. You get amazing speed to lead. Now AI is moving into this thinking level, becoming a strategic advisor in the enrollment space.”
For the counselor on the front line, this looks like AI suggesting which student to call and when. For the leader, it looks like AI surfacing insights from the data that weren’t visible before: what’s worth reporting to the board, what’s shifting in the funnel, what to focus on next.
The shift is decisional, not technological. Early enrollment AI did things for the team. Advisory AI changes what the team chooses to do next. The strategic question moves from what AI tool should we buy to what decisions could an AI advisor sharpen.
The New Search Reality: Students Are Asking AI, Not Google
Sathish: People used to search in Google. Now they start in ChatGPT or Claude. Should universities optimize for that?
Aaron: “Institutions are going to need to change how they position themselves. It’s very common now for students to say, ‘give me a list of all the schools in the western half of the United States that have this major and rank them by cost.’ You need to think like a student and the questions they’ll ask if you want to be on that list.”
Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO) now carry the strategic weight mobile-responsive design held a decade ago. A student building a shortlist with ChatGPT or Claude isn’t browsing program pages. They’re asking one structured question and accepting a structured answer.
That changes what institutional content has to do. Program pages need machine-readable outcomes data. Cost pages need parseable structure. Major listings need to be findable through natural-language queries. Institutions whose content sits behind PDFs and brochure-style layouts are functionally absent from the discovery layer where students now make their first cut.
AI-Informed, But Still Personal
Sathish: Is the enrollment team already using these tools? Anything they should try immediately?
Aaron: “It’s going to change the role of the admissions person. We’ll use AI to rewrite communications flow and make it more responsive. But it’ll also inform the conversations we have: whether now is the right moment to call, whether you’ve built the relationship you need. Every interaction becomes AI-informed.”
Aaron: “We always struggle with speed to lead. By the time we get back to students, sometimes we never get back to them. Students expect quick responses.”
The concrete picture: AI watching a student navigate the website, asking would you like us to give you a call? and placing that call itself.
The bar for follow-up has moved. A student who fills out an inquiry form expects a response in minutes. AI-driven outreach has crossed from competitive advantage into baseline expectation. Institutions still on next-day callback cycles are losing leads they cannot even track. Those leads aren’t bouncing because they were unqualified. They’re bouncing because by the time the call is returned, the student has already moved three institutions further down their list.
Where Enrollment Is Actually Getting Automated
Sathish: Is there a specific part of enrollment that got automated, customer service, financial aid, student support?
All of it is being automated, but the most consequential work is happening in modeling and prediction.
Aaron: “It’s going to have huge impacts on predictive ability, which students are most likely to enroll, what they need to enroll. It’s challenging our assumptions. We’ve had certain assumptions for years that said this interaction is the highest predictor. With more sophistication, you can look behind those impacts and say, maybe that’s not the driver. Maybe that’s the outcome, not the initiative.”
Aaron also emphasizes the analytics speed-up in the last six weeks of a recruitment cycle, when teams need quick adjustments and rarely have time to run them.
Most institutions built recruitment strategies around what seemed to work in past funnels. The next generation of predictive models asks a sharper question: did those interactions drive the enrollment decision, or were they just markers of students already going to enroll? Distinguishing causation from correlation here is the difference between budget that produces yield and budget that produces statistics.
Online vs. In-Person: The Quality Gap Nobody Solves
Sathish: Are students looking for online more than physical campus? Should colleges adapt?
Aaron: “The trend is in that direction. It’s not immediate. There’s still a strong base of people who want that experience of living on campus, interacting in person. We’ve struggled to get the online experience to equal in-person. Almost all institutions need an option in their repertoire. There’s a growing market that only wants education that way.”
“Add an online program” isn’t a strategic move on its own. What differentiates an institution is the quality of the online experience, where AI-driven personalization, tutoring, and student support move the needle. Institutions treating online as a budget version of in-person are competing for the worst-fit slice of a growing market.
Financial Aid: The Slowest Adopter, the Biggest Upside
Sathish: Can financial aid be automated? Could there be a 24/7 financial aid counselor?
Aaron: “Financial aid is sensitive because there are a lot of regulations. Those areas adopt AI more slowly because they’re cautious. Automating packaging has been something that area has struggled with for years. The tools are there, but in most systems I’ve seen, it’s very clunky. It still requires huge upfront effort from financial aid professionals.”
The bigger opportunity sits in student self-service.
Aaron: “So many students get lost in the financial aid process. There’s not enough time for financial aid counselors to call everybody. AI will give students an advocate in the process, a thing that walks in front of you, helps you open the door, helps you talk more intelligently. It might solve the problem, or flag it to a real person.”
Aaron: “Currently you’re not engaging much with families before they’re admitted. If you can predict what they’re likely to need, you can start engaging much earlier, pre-packaging them, helping them know what it’ll look like if admitted.”
The student-facing financial aid experience is one of the highest-leverage automation opportunities in higher ed. Every student who abandons FAFSA, verification, or appeal because the process was confusing is a student who didn’t enroll, and that abandonment stays invisible.
AI as a navigation companion (flagging blockers, nudging next steps, escalating to humans when needed) reclaims abandoned funnels at a fraction of the cost of adding staff. Pre-admission packaging is the leading edge few institutions are doing today.
Why Financial Aid Lags, and Why That Is About to Shift
Sathish: Are regulations the bottleneck slowing this down?
Aaron: “Compliance is much higher in financial aid than in admissions. But part of it is just dynamic. Admissions is always pushing, they want competitive advantage. Financial aid offices don’t have that same dynamic. They’re more a service area. At institutions where they see financial aid as tied to enrollment, they’re more likely to be early adopters.”
Organizational structure is the strongest predictor of AI adoption speed in financial aid. Institutions that treat financial aid as a downstream service desk will lag. Institutions that have integrated financial aid into enrollment strategy will adopt first. If the aid director isn’t already in the AI conversation, that’s a structural problem, not a technology gap.
Retention: The 50% Problem Nobody Wants to Talk About
Sathish: Are there predictive systems to identify “this student is going to have an early exit”?
Aaron: “There have been tools where students opt into a chatbot companion that asks ‘how’s this going?’ Based on feedback, it adjusts, makes suggestions, notifies people. It’s a huge need.”
Aaron: “Nationally, we’re a little over 50% of students getting an undergraduate degree in four years. If you’re only serving about half your customers well, that’s not great. What a huge opportunity in this moment when there are fewer students and more competition. If instead of recruiting 40% more students, you keep more of the ones you have, that’s a huge ROI.”
Sathish: Have you seen tools actually move the needle on retention?
Aaron: “There are examples. It’s not usually a 10% move, usually 2 or 3%. The biggest challenge is there are so many factors, it’s hard to see what’s making the difference. The tools will get more sophisticated.”
| Lever | Typical Cost | Typical Return |
|---|---|---|
| Recruit one additional student | High: marketing, travel, staff time, aid discount | One year of tuition (if they stay) |
| Retain one current student | Low: advising, AI navigation, friction removal | Remaining years of tuition + alumni lifetime value |
Retention is the most undervalued ROI lever in higher ed. A 2 to 3% improvement sustained across cohorts outperforms most marketing budget increases. The student who stays is dramatically cheaper than the one who has to be recruited. AI’s signal advantage is detection: early warning signs (skipped advising, drop-off in learning management system engagement, missed payments) are visible at scale to systems that can flag them in time for a human to intervene.
The AI Companion: Curated Student Journeys
Sathish: Will there be personalized AI tools developed for each university that engage students directly?
Aaron: “AI is going to become that kind of companion. Instead of asking each other, which is a bad source of information, or waiting to ask their advisor, students will ask their AI companion: how do I navigate this? Help me do this.”
The student-facing AI companion will become standard within two enrollment cycles. The question is whether the institution builds it well, tied into its student information system (SIS), financial aid system, and academic record, or whether students rely on generic external tools that don’t know institutional policies. Generic AI confidently gives students wrong answers about specific institutions every day. Institutional AI gets it right.
AI as the Great Equalizer for Smaller Colleges
Sathish: Could AI help smaller colleges compete with larger universities?
Aaron: “If it’s done right, it could be an equalizer. The playing field is more unbalanced now than it has been in my career. If smaller institutions are nimble enough, they could use the technology to provide a highly personalized experience, stronger service, more availability. It’s a great opportunity, but you need to build it specifically to your institution. It’s not one-size-fits-all.”
Smaller institutions have an unrecognized structural advantage. They can move faster on procurement, customize deeper into workflow, and integrate AI into the personal experience large universities struggle to scale. The mistake to avoid is buying off-the-shelf tools designed for large universities and retrofitting them.
Watch the Full Episode on YouTube
Subscribe to the Binary Words channel for more conversations with higher ed leaders.
The International Recruitment Opportunity Nobody Is Taking
Sathish: Is AI making a significant impact on international student recruitment?
Aaron: “International is an area of passion for me. Right now we’re missing a huge opportunity. People are nervous, they’ve seen dips in international enrollment, they’re concerned about the regulatory environment, and they’ve totally backed away. That’s not the right approach. What you do is question the way you’re doing it. Does this still fit in the new environment?”
Aaron points to three shifts that define the opportunity: modality, partnerships, and geography.
Aaron: “I’m working with a couple of institutions on offering online dual enrollment to high school students internationally. Not many people are doing that. That is a huge market.”
Aaron: “I was talking to someone networked with hospitals in a particular country. Those hospitals don’t want to send students for four years. They want a one-year certificate program that helps their employees upskill. That’s a wonderful opportunity. It just requires we think differently.”
Aaron: “Too many institutions became dependent on one or two countries. When things change there, they assume international doesn’t work anymore. I’ve been doing great work in Latin America. That’s a market doing very well.”
International recruitment in 2026 isn’t the 2018 playbook with a new logo. It’s a structurally different opportunity built on three shifts:
- New geographies (Latin America, where regulatory friction is lower and demand is climbing)
- New modalities (online dual enrollment, short-form certificate programs)
- New partners (international hospitals, employers, ministries)
AI’s role here is concrete: market intelligence, language localization, identifying micro-markets where a program would resonate.
Where Should Leaders Actually Invest in AI?
Sathish: If university leaders have to decide on AI investment, where should they put their money first?
Aaron: “It’s a really hard thing as a leader. Oftentimes you feel like you should already know the solution. For a lot of leaders, the first step is becoming more aware of what’s out there. There’s no shame in saying, I need to learn. The best thing many institutions can do is find a partner they trust. You want to build something specific to you.”
Aaron: “It’s an opportunity to ask, how can AI help us find where the leaks are in the boat? Everybody’s dropping the ball somewhere. If our leak is slow follow-up, AI can help. If we don’t know which ones to reach out to, AI can help. If we don’t have visibility into what’s working, AI can help.”
The best decision a higher ed leader can make right now is to not start with the tool. Start with the leak. Map your funnels by revenue impact, then evaluate AI against the largest gaps. Buying technology that solves the wrong problem is the most expensive mistake in higher ed AI today, and most of those purchases trace back to vendor-led discovery instead of leak-led discovery.
Hallucinations, Risk Tiers, and Where Human Checks Matter
Sathish: Hallucinations are always a concern. Students might make decisions based on AI responses. How should universities think about governance?
Aaron: “It’s easy to forget that problem exists. You just get used to relying on it. Always put in human checks and balances. If you’re moving ahead with AI in financial aid, you better have somebody quality controlling it, precision is critical. In other areas, maybe it isn’t. If you ask AI ‘which students should I call today,’ and it gives you a couple of wrong ones, you called them extra, not a big deal. Where the risk is low, let it loose. Where it’s larger, be on the lookout.”
AI governance in higher ed isn’t one policy. It’s a risk-tier framework:
- Low-stakes uses (outreach lists, draft communications, meeting summaries) can run with lighter oversight.
- High-stakes uses (financial aid packaging, regulatory reporting, anything affecting a student’s official record) need explicit human checkpoints and audit trails.
One policy stretched across both either over-restricts low-risk uses or under-protects high-risk ones, which is where the liability lives.
Will AI Replace Enrollment Jobs?
Sathish: Is there fear that AI is going to impact jobs in enrollment?
Aaron: “I think it will shift jobs. I don’t think it will eliminate as a quantity. Compare it to when email first appeared. We all thought, I’m going to have so much less work. Now we have lots of email. You saw shifts, fewer secretaries, no one taking dictation. But it created jobs in other areas, because the technology made new things possible.”
Aaron: “You’ll need fewer people to do resume reviews because AI can write the resume just as well. But you’ll need more people to teach students soft skills, how to network well, how to think strategically about your career long-term. The human interactions you do have will become much more important and need to be much higher quality.”
The jobs at risk are the ones built around volume tasks AI now handles well: first-draft writing, basic data review, repetitive outreach, routine resume edits. The jobs that grow require judgment, relationship depth, and strategic thinking. Reskilling enrollment teams isn’t a 2028 problem. It’s a 2026 problem.
The 3 to 5 Year Outlook: From Stylistic to Structural
Sathish: In the next three to five years, how will enrollment and universities look?
Aaron: “Higher ed has been waiting for something to change to get unstuck. We’ve been stuck in old models. In three years, you’ll see stylistic changes, handling applications differently, highly personalizing things. But on the five-year range, it may make dramatic shifts. The whole application process might change because so much of what students have been turning in can now be done by AI. Do you need to still collect it? It could be transformational.”
Three-year planning is about optimization. Five-year planning is about reinvention. Most institutions are planning on the three-year horizon, optimizing processes that may not survive five years at all. Institutions setting the next standard are already asking what the application, the financial aid letter, and the recruitment funnel should look like in 2031.
The Hidden Revenue Crisis: Why the Old Model Is Broken
Sathish: What is the biggest opportunity or biggest pain point in enrollment and financial aid right now?
Aaron: “Students are fleeing to the very highest-level institutions because there’s concern about how degrees translate into job opportunities. Students chase those few with the biggest brand names, thinking that’s where they’ll secure their future. The data doesn’t suggest they have any better outcomes that way. The quality of the experience you have at the institution is much more predictive of job success than the selectivity of where you go.”
Aaron: “Three quarters of US institutions, maybe 80%, are not those highly selective institutions. There’s always been this idea, probably from the 60s or 70s, that every year you recruit more students and raise tuition slightly. That’s the model. What happens when there aren’t more students to recruit, and the ones who exist are less willing to pay tuition increases? You’ve broken that financial model. You’re seeing budget cuts, institutions unwilling to invest even when they need it most.”
This is the crisis hiding underneath every other higher ed conversation. The “more students plus small tuition increases” model that drove planning for fifty years no longer fits the demographic or economic environment. Institutions trying to AI-optimize the old model are speeding up the path to a dead end. The real move is using AI not to optimize the old revenue model but to redesign it: new markets, new modalities, new partnerships, new ways of capturing value institutions have been writing off for decades.
Finding the Leaks: A $5.6 Million Real-World Example
Aaron: “Every meeting at Mac Insight Group, we talk about wanting to be the ROI company for institutions. We’re not satisfied with two to three times ROI, we want 10X. That’s how institutions need to think: how does AI help us rethink our revenue streams?”
Real-World Result
$5.6 million recovered in pre-collections over two years.
“We worked with a client, an institution in California. In two years, we helped them recoup $5.6 million in pre-collections. Before sending a student to collections because they hadn’t paid their bill, we reach out, talk to them, give them options, and they pay part of the bill. $5.6 million. People don’t even think about that. AI will help us find those opportunities. Your boat is leaking, here’s where it is. You just lost six million dollars because you didn’t help people pay their bills.”
Aaron: “We’ve been stuck in this mentality that it’s about more students. In an environment where there aren’t more students, you have to shift mentality and say, this is about revenue. The more AI can point out those obvious problems, the more it’ll help us transform.”
Revenue in higher ed isn’t just new students. It’s:
- Unpaid bills written off
- Leads never called back
- International markets quietly abandoned
- Students who dropped out in year two because nobody noticed
Each is recoverable revenue, and AI is increasingly the tool that surfaces those numbers in time to act. The institutions that win the next five years aren’t the ones with the largest enrollment growth. They’re the ones that stopped letting six and seven-figure leaks run silently down the side of the boat.
Final Advice: Opportunity, Not Threat
Sathish: Any advice for future enrollment leaders navigating this shift?
Aaron: “See it as an opportunity. It’s easy to see it as threatening: it’s going to hurt jobs, make the process less personalized. What’s harder is what we don’t know is going to happen. There’s a way to use AI as your companion and say, where am I missing? Use my gut instinct, my training, my best people. Reinvest in relationships. This is a way to scan your environment and see where the issues are. It’s not about trying harder. It’s about trying differently.”
Aaron’s final framing reframes the entire conversation. AI doesn’t fix the boat. It tells the institution where the water is coming in. The fix is still on the institution, but for the first time, the leak is finally visible in time to act on it.
Action Checklist for Higher-Ed Leaders
AI is already reshaping how students find, choose, and complete their education. Institutions that act on specific leaks now will be positioned to lead. Those that wait will face harder choices later.
Discovery and Audit
- Map enrollment, financial aid, and retention funnels by revenue impact, not by department. Most leaks sit in the gaps between departments. Revenue-based mapping reveals them.
- Identify your three largest revenue leaks. Slow follow-up, FAFSA abandonment, year-two drop-off, and pre-collections write-offs are the most common. Know which ones you’re losing to.
- Audit program, cost, and major pages for AI-readable structure. If your content lives behind PDFs or brochure-style pages, you are invisible to GEO and AEO search.
Enrollment and Outreach
- Measure current speed to lead in minutes. If you cannot answer this in minutes rather than days, you are already behind baseline expectation.
- Pilot AI-informed outreach: which students, when, and why. Start with a small cohort. Let the data challenge your existing assumptions about which interactions actually drive yield.
- Add AI-driven on-site engagement on your highest-traffic pages. Program, cost, and financial aid pages are where intent converts or abandons. That is where real-time engagement matters most.
Financial Aid
- Bring the financial aid director into the AI conversation now. If they are not already at the table, that is the first structural problem to fix.
- Track FAFSA, verification, and appeal abandonment rates. Every abandonment is a student who didn’t enroll, and most institutions have no visibility into where in the process they left.
- Pilot pre-admission packaging for the next admit cycle. If you can predict what a prospective family needs before admission, start that conversation early. It is one of the highest-leverage moves available.
- Build a pre-collections outreach workflow. Reach students before the bill hits collections. The $5.6 million recovery Aaron describes started with this single change.
Retention and Student Experience
- Set a 2 to 3% retention improvement target for next year. This modest lift sustained across cohorts outperforms most marketing budget increases on ROI.
- Track early warning signals: skipped advising, LMS drop-off, missed payments. These signals are visible at scale. You need systems that surface them in time for a human to intervene.
- Decide whether to build an institutional AI companion or accept generic AI. Students will use AI to navigate their academic journey regardless. The question is whether that AI knows your institution’s actual policies.
International
- Evaluate Latin America as a recruitment market. Regulatory friction is lower and demand is climbing. Institutions that pulled back from international entirely are missing this.
- Explore online dual enrollment for international high schoolers. Very few institutions are doing this. It is a large, underdeveloped market.
- Look at employer or hospital partnerships for certificate programs. Short-form, employer-aligned credentials for international partners represent a structurally different revenue stream that many institutions overlook entirely.
Governance and Team
- Build a risk-tier AI governance framework. One policy stretched across low-stakes and high-stakes uses simultaneously either over-restricts or under-protects. Tier your oversight by consequence.
- Define human checkpoints for high-stakes uses. Financial aid packaging, regulatory reporting, and anything affecting a student’s official record need explicit review steps and audit trails.
- Begin enrollment-team reskilling now. Volume tasks are being automated. The roles that grow require judgment, relationship depth, and strategic thinking. This transition is already underway.
Strategy
- Replace vendor-led with leak-led discovery in your next AI procurement. Start with the leak, not the tool. Buying technology that solves the wrong problem is the most expensive AI mistake in higher ed today.
- Build a five-year reinvention plan alongside your three-year optimization plan. Most institutions are optimizing the three-year horizon. Processes you are optimizing today may not exist in five years. Plan for both.
Ready to Find the Leaks in Your Own Boat?
If this made you think about where your institution might be quietly losing revenue, missed leads, abandoned financial aid journeys, retention drop-off, or international markets pulled back from, let’s talk. BinaryWorks helps higher education institutions identify where AI can meaningfully sharpen enrollment, financial aid, and student experience strategy. Built specifically to your funnel, your aid model, your student.
- • AI-enhanced digital platforms built for enrollment readiness and governance
- • GEO and AEO content strategy to make your institution visible in AI-native search
- • Funnel analysis and leak identification tied to measurable revenue outcomes
No pitch. Just a real conversation about where AI fits in your enrollment and revenue strategy.
Keep the Conversation Going
This is AI Talks #4 by BinaryWorks, inside stories and strategies from leaders navigating the shift to AI, automation, and the digital infrastructure they require.
Also Watch: AI Talks #3 | The Human Side of AI: Leading with Purpose in Higher Education →