June 16, 2026 · Episode 316
When AI Finds the Administrative Friction Higher Ed Leaders Miss
36 Min · By Dr. Drumm McNaughton
Administrative AI helps higher ed leaders find friction in enrollment, retention, student support, and advancement before it costs revenue.
A 4% year-over-year retention gain sounds like a service-desk metric until the dollars get attached to it. At a 5,000-student institution charging $25,000 a year, with roughly 1,500 students per class, that lift compounds across the years a student stays enrolled into real money. Empire State University reported exactly that gain after rolling out AI-based virtual assistants across roughly eight departments, alongside a 25% increase in student engagement on those web properties. The correlation is meaningful for leaders building a business case, though leaders should treat it as a correlation worth investigating rather than a clean causation claim.
Administrative AI is producing measurable financial gains at institutions that have paired implementation with strong change management. Empire State, Florida Southwestern, and Boise State are not the largest or best-funded universities in the country, and each has numbers that would change a board conversation. The institutions treating administrative AI as a future decision are losing ground every cycle to the ones treating it as an operating priority now.
Justin Beck, CEO of Gravyty, works where higher education’s administrative systems meet student and constituent engagement. Gravyty connects with the CRM, ERP, and SIS platforms that support enrollment, retention, advancement, and donor outreach. His central point is direct: AI is already present on campus, but its greatest near-term value may be in finding the administrative friction leaders cannot see until it affects enrollment, retention, student support, or donor revenue.
Retention Is the Strongest Financial Case for Administrative AI
The Empire State numbers, reported by Gravyty, are worth holding against any business-case spreadsheet. Engagement across roughly eight departments moved 25% after AI virtual assistants went live. Retention moved 4% year over year. At a typical mid-size institution, four percent of retained tuition revenue, held across the years a student stays enrolled, runs into the millions. Leaders building a business case should model the gains rather than accept them at face value, but the magnitude is what makes administrative AI worth the modeling.
The student case sits underneath the institutional one. Students who do not complete often carry loans without the credential to pay them back. Higher ed’s threefold purpose, as Dr. McNaughton frames it on the show, is workforce preparation, the public good, and the transformation of a teenager into an adult. Debt without a degree undermines all three. Administrative AI that closes the gap between enrollment and completion is doing more than protecting tuition revenue. It is reducing the population of borrowers left with debt and no credential.
Student AI Adoption Is Settled. Institutional Governance Is the Open Variable.
A Cal State system survey of more than 94,000 students, faculty, and staff found 95% of respondents had used at least one AI tool, with more than half of students reporting regular use. Roughly 80% of student respondents said they are not comfortable submitting AI-generated work as their own. The takeaway for leaders is not that AI use is widespread; that point is settled. The takeaway is that institutional governance, not student behavior, is the open variable.
The Real Implementation Risk: Bots That Behave Like Call-Center Bots
Every leader who has called a company recently knows what a bad bot does. It loops, refuses to escalate, and burns the relationship. Beck readily admits he hits zero on customer service calls himself once a bot starts looping. Dr. McNaughton’s rule is harder: three failures and the product loses his recommendation entirely. An admissions chatbot that behaves the same way is not a neutral tool. It is an enrollment liability.
The previous generation of admissions chatbots ran on scripted decision trees with a hard human handoff the moment a question went off-path. The current generation, in Beck’s framing, blends a defined knowledge base for predictable questions with a guardrailed generative layer for the off-path ones. That blend can reduce looping when paired with a defined knowledge base and clear escalation rules, because the generative layer can answer questions the script writer did not anticipate.
This is where the Florida Southwestern numbers come from. Gravyty reports that the college resolved approximately 90% of inbound admissions inquiries on first contact after deploying AI tools across recruiting and admissions, with time to class registration dropping by half. Beck also cited Indiana University absorbing roughly 40,000 additional applications in a recent cycle with the same staff, an example of how AI sort-and-triage can carry volume that hiring cannot keep up with. Institutions without AI sort and triage face a choice between hiring they cannot afford and yield they cannot recover.
The infrastructure around the bot matters as much as the bot itself. Live analytics need to surface which questions are stumping the system and where students are abandoning workflows. Reviewed daily rather than quarterly, that data is what lets the institution intervene before a fixable problem becomes an enrollment problem.
AI Reveals Where the Institution Is Misaligned to Itself
At one New York institution Beck has worked with, more than 40% of questions coming into the financial aid office had nothing to do with financial aid. They were housing questions, registration questions, anything students could not get a clear answer to elsewhere. Financial aid was not overloaded because financial aid was complicated. It was overloaded because the rest of campus had information fragmented across enough systems that students defaulted to the office most likely to help.
A generative layer pulling from the full campus knowledge base gives the student a usable answer and shows the institution where its own systems are disconnected. That cross-departmental view is almost impossible to produce any other way. Each office runs its own ticketing system with its own backlog, and the patterns across them only become visible when a single AI layer can see them at once.
The diagnostic application extends beyond inbound questions. An AI web crawler at one Missouri institution surfaced three different admissions deposit dates published on three active web pages. The institution updated the pages within an afternoon. Without the scan, Beck noted, the institution would not have caught the inconsistency until deposits arrived late or did not arrive at all.
Advancement AI’s Real Value Is Donor Prioritization, Not Email Generation
Advancement has been slower than enrollment to adopt AI, in part because of legacy systems and entrenched workflows. Boise State University moved early. Per Gravyty’s published case study, an advancement AI rollout gave major gift officers a prioritization layer for portfolios of 150 to 200 prospects, drafted personalized outreach, surfaced analytics on each donor’s signals, and powered curated digital experiences across the donor base. Beck describes the program as a combination of those elements rather than any single feature.
The reported outcomes set a useful benchmark for leaders to investigate against their own baseline. An 87% year-over-year increase in donor volume. A 50% increase in donor interaction and response. $635,679 raised through the AI-assisted channel. The contribution that is hardest to replicate without AI is the prioritization across a 150-to-200-prospect portfolio. Knowing which donor to contact, when, and with what context is the work major gift officers have historically done in their heads, at the limit of how many relationships any one person can hold.
The fundraising principle underneath this is well-known and rarely operationalized at scale. The first interaction with a potential donor is not an ask. AI’s payoff in advancement is the long stewardship cycle, including a use case that is still emerging: matching current students with alumni mentors based on interests and skills. Affinity built before graduation produces alumni who give later, and matching across populations of tens of thousands of students and alumni is hard to operationalize at that scale without automation.
Half an Hour a Week Is What Separates Working AI From Failing AI
Administrative AI is not a set-it-and-forget-it deployment. Beck’s exact framing: no Staples Easy Button, nothing happens auto-magically. A virtual assistant supporting financial aid or admissions needs roughly half an hour of human attention per week, in Beck’s estimation. The review looks at the top 20 questions coming in, when they are arriving, what students do next after the bot responds, and which questions are abandoning the workflow before resolution. Without that cadence, the bot drifts out of alignment with how the institution actually operates.
The same review is where AI shifts from a response engine to an instrument for continuous improvement. Outdated FAQs surface. Misrouted queries surface. Stale knowledge in policy-driven domains, where federal financial aid rules and Title IX guidance and accreditation requirements all change on their own schedules, gets updated continuously rather than annually. Beck estimates the work this replaces at roughly two FTEs of policy-update labor, which most institutions do not have to spare.
Acceptable Use Policy First, Private Instances Second
An acceptable use policy is the precondition for everything else. Beck estimates well over half of institutions have one in place and the rest are writing them. Without one, there is no liability protection, no community guidance, and no defense against individual departments procuring their own AI tools in ways that fragment data and expose information that should not leave the institution.
A parallel trend is reshaping where institutional data lives. More institutions, particularly those in the top Carnegie classifications, are standing up private instances of AI tools so proprietary data stays inside the walls. Cost makes the option uneven across the sector. The principle does not have to be. The institution itself should be sponsoring the models and centralizing the tools rather than letting departments run wild with their own procurement.
Three Takeaways for University Presidents and Boards
- Move quickly, but with the structure in place. An acceptable use policy on the books. Defined institutional outcomes the AI work is supposed to drive, framed in the language the institution already uses: student success, enrollment success, and alumni success. Fast enough that AI use does not sprawl across departments in ways that have to be wrangled back in later.
- Control what you can control, but keep your stakeholders in mind. Pull them in. This is one of the strongest opportunities for the faculty committee model, including what Beck called his least favorite term in the world until he started doing AI work. Bring in the full spectrum, from skeptical newcomers to experienced practitioners.
- Iterate often. The technology is changing month by month. Establish a governance structure that brings the right people back together on a recurring cadence to review whether the programs are still effective and where they need to pivot.
Beck compared the trajectory to NIL in athletics. The NCAA took a long time to get its arms around NIL, which left every institution to solve it for itself, which meant nobody really solved it. Administrative AI is heading down the same road. The institutions still treating it as a debate are accumulating opportunity cost on every cycle: retention dollars not recovered, donors not re-engaged, applications not processed, students not retained. Administrative AI is becoming part of higher education’s operating infrastructure, and the institutions that build the governance and implementation capacity now will set the benchmark the rest of the sector measures itself against.
For leaders who want the implementation specifics behind the numbers above, the full episode goes deeper: how the canned-content and generative-AI layers actually work together inside a financial aid deployment, how the Boise State prioritization model was structured against a 150-to-200-prospect MGO portfolio, and where Beck has watched institutions go wrong on the rollout sequence.
Quick Overview
- A 4% year-over-year retention gain at a 5,000-student institution charging $25,000 a year compounds into a multi-million-dollar financial case for administrative AI.
- Empire State University reported that 4% retention gain alongside a 25% engagement increase after deploying AI virtual assistants across roughly eight departments, per Gravyty’s published outcomes.
- A Cal State system survey of 94,000+ respondents found 95% had used at least one AI tool, with more than half of students reporting regular use and roughly 80% uncomfortable submitting AI-generated work as their own.
- Florida Southwestern State College resolved approximately 90% of inbound admissions inquiries on first contact and cut time to class registration in half, per Gravyty’s published outcomes.
- Over 40% of questions coming into one New York institution’s financial aid office had nothing to do with financial aid, exposing fragmented student-facing information across the institution.
- Boise State University reported an 87% increase in donor volume, a 50% increase in donor interaction and response, and $635,679 raised through the AI-assisted channel, per Gravyty’s published case study.
- AI-supported mentor matching between current students and alumni builds donor affinity years before any ask, and is hard to operationalize at scale without automation.
- About half an hour of human review per week keeps a virtual assistant aligned with institutional operations and replaces what Beck estimates at roughly two FTEs of policy-update labor.
- The technology is changing month by month, which requires a recurring governance cadence rather than a one-time decision.
About Our Guest
Justin Beck is the CEO of Gravyty, a leading provider of AI-driven engagement solutions for higher education and nonprofit institutions. With over two decades of experience in SaaS and education technology, Justin has built and led high-performing global teams across sales, customer success, revenue operations, marketing, and product enablement. He is driven by a deep belief in the power of technology to improve outcomes across the student and constituent lifecycle—from recruitment and retention to alumni engagement and fundraising. Prior to Gravyty, he served as Chief Commercial Officer at Xplor Technologies, a global, multi-vertical SaaS and embedded payments company serving over 100,000 customers in over 30 markets. Justin’s earlier leadership roles include SVP of Global Sales at Instructure (makers of Canvas), and senior positions at Salesforce, Kaltura, EverFi, Blackboard, and Apple. Across each organization, he has played a pivotal role in scaling revenue, entering new markets, and driving strategic transformation.
About the Host
Dr. Drumm McNaughton is the founder, CEO, and Principal Consultant at The Change Leader, Inc. A highly sought-after higher education consultant with 20+ years of experience, Dr. McNaughton works with leadership, management, and boards of both U.S. and international institutions. His expertise spans key areas, including accreditation, governance, strategic planning, presidential onboarding, mergers, acquisitions, and strategic alliances. Dr. McNaughton’s approach combines a holistic methodology with a deep understanding of the contemporary and evolving challenges facing higher education institutions worldwide to ensure his clients succeed in their mission.
Read the Podcast Transcript →
Transcript: Changing Higher Ed podcast 316 – with host Dr. Drumm McNaughton and guest Justin Beck
Introduction to Changing Higher Ed®
Welcome to Changing Higher Ed®, a podcast dedicated to helping higher education leaders improve their institutions. With your host, Dr. Drumm McNaughton, CEO of The Change Leader, a consultancy that helps higher ed leaders holistically transform their institutions. Learn more at changinghighered.com. And now, here’s your host, Drumm McNaughton.
[00:00:31] Welcome and Guest Intro
Drumm McNaughton: Thank you, David.
My guest today is Justin Beck, CEO of Gravyty, a leading provider of AI-driven engagement solutions for higher education and nonprofit institutions. Justin has a deep belief in the power of technology to improve outcomes across the student and constituent life cycle, from recruiting and retention to alumni engagement and to fundraising.
He says that higher ed needs to stop asking whether or not AI belongs on campus and focus on the reality. AI is here, and institutions need to be implementing it in ways that can deliver true value across campus. Justin joins me today to talk about the administrative applications of AI and how it can improve institutions, their effectiveness, their recruiting, their enrollment, and their retention.
Justin, welcome to the program.
Justin Beck: Drumm, it is wonderful to be here. Humbled to be in your presence, and I appreciate the invitation today.
Drumm McNaughton: Well, flattery will get you everywhere, my friend. So I’m looking forward to a great conversation.
[00:01:42] Gravyty and AI Focus
Drumm McNaughton: You lead an AI-driven engagement solutions company called Gravyty.
Justin Beck: With a weird spelling since every great name on the planet’s already taken, you have to like finagle with the letters sometimes. But yes, the company’s name is Gravyty.
Drumm McNaughton: Understand. And you’ve got an incredible background in technology, Canvas, other places as well. Give us a little bit about your background. We’re going to talk about AI specifically. This is going to be the fifth in a series of AI podcasts, but we’re going to talk about something different. This is an area that you excel in, is the implementation of AI across administrations. So give us a little background, please.
Justin Beck: Yes.
[00:02:26] Justin’s EdTech Journey
Justin Beck: So quick personal background. I don’t like talking about myself. It’s, the Catholic Midwest kid in me makes it really tough to talk about oneself. But cut my teeth and grown up in, like, the educational technology space. Early days at Blackboard, which was one of the early leaders of online learning pioneer. Got to see that adoption.
Drumm McNaughton: I remember it.
Justin Beck: Yes, it’s a great seminal time period of life, really seeing the modalities of education change. And then career progressed into various go-to-market roles at, you mentioned, Instructure, global head of sales there. Led a company called Kaltura, which is a video streaming software organization that focused on commercial organizations as well as higher education globally. Great growth and experience and customer impact stays there.
And then, as you shared, find myself at the helm of an organization called Gravyty, which is?
[00:03:12] What Gravyty Actually Does
Justin Beck: When you think about Gravyty, we sit at the intersection of where systems of record, like the CRM systems, the ERP, the SIS systems that power fundraising and student engagement and core records of university. We bring personalized action at scale. So I know those are buzzy words, but we’ll talk about the nuances and specifics, how those manifest. But ultimately, it’s about how do you take the data that’s living inside your CRM and your SIS and your ERP systems and make them actionable through engagement across like a full student life cycle.
Drumm McNaughton: You mean we’re going to do away with those APIs?
Justin Beck: Those APIs are never, go-we’re never going to get away from those APIs. They’re the lifeblood by which we’re able to push data back and forth between those host systems and organizations like ourselves.
[00:04:01] APIs Explained Simply
Drumm McNaughton: And of course, I just jumped into it right off the bat by saying API. What does API stand for? because most of the folks who listen to the podcast are academics. They may not know what an API is and what an API does.
Justin Beck: Yes, an API, when you think about it, it’s, it’s basically like the pipes by which data can move in and out of systems, and it’s like a standardized format where you can interface. Early pioneers of APIs were companies like Salesforce and the HubSpots of the world that really made it easy for third-party tools to integrate, but also for developers to take under a common language in a safe, secure, reliable way, move data in and out of their systems and platforms.
Drumm McNaughton: So it’s an application interface. In essence.
Justin Beck: Correct. Yes, I was trying to avoid the techie, but you’re smarter than me, Drumm, so I’m going to let you—
Drumm McNaughton: Oh, horse feathers.
Justin Beck: —handle the techie lingo address. Yes.
Drumm McNaughton: Horse feathers. I just got stuck in that world a little bit years ago when I was with Dell. Gravyty, Gravyty is interesting.
[00:05:06] AI Is Already on Campus
Drumm McNaughton: In talking with you before, you made a quote that really resonated with me is that, “higher education needs to stop asking whether or not AI belongs on campus and focus on the reality. It’s here, and institutions need to be implementing it in ways that deliver value across campus”.
Justin Beck: Yes, it’s on two fronts there.
[00:05:29] Consumer AI in Student Life
Justin Beck: One, in that there’s obviously a consumer, and when we think consumer, you think student and faculty members to some extent. This has become commonplace in all of our world, between OpenAI, Claude, and all the various large language models. Folks are using that today now more than ever, and I think that goes without saying.
So it’s natural that it starts extending into your professional or education life, that you start using these tools for your own, personal productivity and your personal outcome. That intrinsically then prompts the how do I make my education experience, whether that’s content creation as a faculty member, content assessment, or as a student.
As my teenage children remind me all the time, they’re using multiple AI tools to help with their writing, refinement, voice language testing, quizzing. It’s not all deviant , as sometimes it gets hyperbolized. Like my daughter is a high schooler and she was using AI to create flashcards for herself on a subject matter that she was saying.
Now, she could have gone and procured a set of flashcards, but she was using OpenAI to create flashcards for her. So, I think one from that consumer perspective.
[00:06:34] Institutional AI and Governance
Justin Beck: And then the secondary piece is that institutions, your systems that they engage with, whether again, there’s ERPs, SIS, and various ed– academic and administrative tools, are all putting AI functionality into their product platforms today.
So to your point, Drumm, it’s here. Now it’s about how do we manage it? How do we keep it secure? How do we implement properly? And how do we look for scale opportunities where we can really drive institutional outcomes, whether that’s student success, student retention, faculty success, faculty retention, administrative optimization.
There’s so many areas that this is presenting an opportunity for us to bunk into.
Drumm McNaughton: Mm-hmm. And it’s interesting because you referred to students and faculty using it. Cal State did a survey, over 90,000 respondents, 90,000, and they said 80 to 90% of all people are using, students especially, are using AI right now, and 80% of them felt like there are unethical uses of it.
[00:07:40] Admin AI Use Cases Overview
Drumm McNaughton: We’re not going to talk about that so much now, but we’re going to start getting into the administrative perspective versus the academic.
We’ve talked about the academic perspective quite a bit in previous podcasts. Where you guys are excelling at is that administrative perspective. Tell us about that.
Justin Beck: Yes, so where Gravyty comes into play, and there’s other organizations like us above, of course, but I think this is a scale and a reach opportunity, is that our organization focused on, think about like student life cycle services as well as engagement with alumni and on the fundraising side of things.
So, one of the key areas where we really interface with institutions is helping with the both admissions and enrollment optimization, as well as leveraging those AI t-powered tools for student engagement and support. So think about an agent or a bot or a virtual assistant handling inbound student questions, no matter the time of the day, around financial aid, around housing, around registration, providing proactive nudges to those students that, “Hey, you’re stuck in a part of this funnel.”
And schools talk about their students as a funnel, as they join in. So proactive nudges for reminders, encouragement, next steps, with the goal of ultimately higher enrollment yields, improved retention, and ultimately reducing the burden on the staff, where if you know anything about higher education right now, they have not expanded their staffing to keep up with enrollment ex- increases or declines, that, that may be happening.
So reducing the load on the staff members is employee satisfaction and better outcomes from them engaging with the community as a whole.
Drumm McNaughton: So there’s really three areas that we’ve talked about that is ripe for use of AI. We just talked a little bit, and we’ll go into more depth on recruiting and admissions. But there’s also the IT and student support and alumni relations and donor stewardship, including athletics.
[00:09:45] Recruiting and Admissions AI
Drumm McNaughton: Let’s dive down into each of those areas a bit, starting with the recruiting and admissions.
Justin Beck: Sure.
Drumm McNaughton: You just made a great point that many institutions haven’t upped their enrollment, their staffing for enrollment. It makes good sense, especially, for many institutions, your flagship state institutions, your Ivys, your top tier are increasing enrollment, but there’s many others that aren’t.
How can AI help drive greater enrollment?
Justin Beck: Yes, I think there’s two parts.
[00:10:19] Case Study Faster Enrollment
Justin Beck: And Florida Southwestern State College is a great example that jumps out to me as one where the AI tools, ours as well as others, help them address some ninety percent of their inbound inquiries were able to be resolved in first contact.
You know, when you think about ninety percent, so if someone reaches out with an inquiry or question, ninety percent of those get answered, and then they’re able to hand off the remaining ten percent for someone to get actionable about. So you think about a staff member in the recruiting and admissions department answering all these rote, redundant questions.
The scale and the technology and the AI-powered capabilities allows them to focus on extreme cases as well as folks that need a little more hand-holding in the process, so that allows individual scale. What that ultimately created was 2x quicker time for folks to get registered for their class, which for the university, that’s dollars.
Drumm McNaughton: That’s money in the bank.
Justin Beck: That’s money in the bank for an organization like that.
What I like about that’s Florida Southwestern, that’s kind of in that smaller institution that you were mentioning that typically do not have as much resources to be able to throw bodies, energy, a bunch of technology at it. Now, on the top end, though, you think about Indiana University or some of the large community colleges in California, like you mentioned, those folks have potentially a few more resources at their disposal, but they’re getting heavy traffic.
We’re talking forty… Like Indiana University, I believe, reported about forty thousand more applicants this last go around.
Drumm McNaughton: Wow!
Justin Beck: How do you handle that volume when you have the same staff, the same systems, everything in place? So AI helps you sort, prioritize, and be able to focus on the work that’s going to be most important on that recruiting admissions side of things.
Drumm McNaughton: That is amazing, 40,000 more inquiries than before. I can only imagine what that would be like from a staffing perspective if you had to have bodies doing that.
[00:12:21] Avoiding Bad Chatbot Experiences
Drumm McNaughton: But I’ve got a question for you, and nobody’s been able to answer this for me well enough, but I have no doubt that you’ll be able to.
Justin Beck: You’re me up, Drumm. You’re setting me up.
Drumm McNaughton: How many times have you called and gotten a bot that wants to steer you down a certain path, then it is so inflexible that you can’t get down that path, and you just end up hanging up the phone? That to me seems like a real problem of what it could be for colleges and universities using AI as the first touch.
How do you get around this?
Justin Beck: Drumm, that’s a great question. What my mind jumps to and where we get to work with a lot of institutional partners is there’s a process flow that obviously you’re trying to design on the support side of things, and this comes back to that implementation question you talked about.
It’s like, “how do we implement these tools so they can be most effective?” Three years ago, the best thing at your disposal was, hey, we’re going to design a workflow, we’re going to give it proper questions and response, and anything that deviates outside of that, you have to flip over to a live person. This is where, with the continued advancement of generative models and all the large language models, is that now there’s this beautiful intersection between the canned content, frankly, that we want to deliver someone in, the perfect scenario, and when they fall off that path, we have this guardrailed generative response that helps them work through in a very natural way what they’re trying to get to before trying to maybe flip them to, to talk to an individual person.
We call it the blending of your dedicated knowledge sources with generative AI starting to come together to allow that information seeker or that student or that faculty member that has a question about HR operations at the university to get a more productive response and a better response in a natural kind of way.
So hopefully that’s a good, a good answer to a very challenging question.
Drumm McNaughton: Yes, well, thank you. It does answer it to some degree, but it also tells me that we’re having to rely on AI more we did to come up with appropriate answers, and then get AI to think when we’ve played Stump the Chump and we’ve gotten a question there that AI can’t answer. It seems like with corporations, they try and steer you down that path, and you’ve got to go down that path three times. I can guarantee that I’ve got a lot of patience. My wife tells me all this all the time. But somebody, at that age trying to apply, are they going to have the patience that, that’s necessary?
[00:15:09] Analytics to Fix Drop Offs
Justin Beck: No, they’re going to abandon rather rapidly, and this is where the curation of those experiences and the analytics that come out of tools like ours and tools that, that, that are available is so important because you can see with exact precision where students are getting stumped on questions, where they’re abandoning a workflow and a process, which allows you r- in real time to take action.
This isn’t a quarterly report that comes out in some packet. You can see live time every day that, boy, if somebody… one of the classic examples where I talk about oftentimes is many times we have an institution, in, New York where over 40% of the questions that come to their financial aid office have nothing to do with financial aid.
Oftentimes they’re about housing, they’re about registration, and if the knowledge being provided to that student, that virtual experience, is just about financial aid, they’re going to get very frustrated and have to go searching somewhere else. Compared to the generative capabilities that can look at all the campus’ knowledge resources or best can get them to the right direction, but also allow the financial aid office to pull in more information to their question and response opportunity.
So that’s where data analytics and active human intervention to curate these AI experiences is so paramount to the success.
Drumm McNaughton: Yes, that, that makes perfect sense. Thank you. You’re a lot smarter than you said you were.
So driving greater enrollment using AI at the top of the funnel is very important, and there’s models to do that.
[00:16:49] Preventing Summer Melt
Drumm McNaughton: One of the things that you said when we spoke the other day was you can use AI to avoid the summer melt. Tell me about that.
Justin Beck: Yes, I’m thinking through some examples of some recent organizations we’ve talked to. So there is obviously a automated student support process where AI can have a great help, too. So there are triggers that happen inside the CRM system, the registration system, the campus housing system.
Think of all the tools that a university has at its disposal, but oftentimes, even in today in 2026, those are fragmented silos of data. So aggregating tools like Gravyty or some of the reporting tools that might be outmark give you triggers and indicators proactively and help funnel down the students that need some intervention, that either hasn’t registered for that upcoming semester, haven’t completed so-some set process.
So very much like the recruiting and admissions action and intelligence, you get very similar indicators on the student success side of things when you have a very clear outcome-driven, prioritization of what you’re hoping, like a bot or a analytics tool is ultimately being helpful for you all.
[00:18:04] Retention Gains and ROI Math
Justin Beck: So Empire State University is a great example where they through implementing aI-based virtual assistants across the numerous departments, upwards of eight departments, they saw a twenty-five percent increase in student engagement with those web properties, which, which, oh, I would love for us to take claim for all this, but ultimately it led to a four percent increase in their year-over-year student retention.
Drumm McNaughton: Mm-hmm. And equate that to dollars so people understand that’s big.
Justin Beck: Yes. So if you think about, an institution that has five thousand students, fifteen hundred of them in any given class, four percent, twenty-five thousand dollars a year, the math adds up pretty darn quick over a three to four-year time period of, of revenue staying with the organization.
And at the end of the day, Drumm, you and I both know the revenue to the organization is super important.
[00:18:58] Why Retention Matters for Students
Justin Beck: But where I think all of us need to be paying attention is these are students who oftentimes have loans, and if they’re not getting to degree completion and productivity in the workforce, we’re saddling a generation with just extreme loans without employability.
And that’s the stuff I know I tend to talk a lot about enrollment funnels and dollars to the institution, which is, ultimately one of, one of our core customers. But this is all truly about making sure students are set up for a lifetime of success.
Drumm McNaughton: Which goes back to what many folks have heard me say, there’s a threefold purpose for higher education: prepare people for the workforce, for the public good, and in many respects, most importantly, the transformation of the student from a teenager to an adult. And with all those things, if they’re saddled with this debt, how can they get out and be more productive when they’re having to pay back things?
I mean, you can’t live a life that way.
Justin Beck: No, particularly with the cost of everything unfolding from housing to… yes, it’s so present and it’s one of the reasons I think I’ve stayed working in the educational technology sector for so long is that it’s so tightly tied to generational outcomes improvements for our students and their families and generations to come.
Drumm McNaughton: Yep. So we’ve covered recruiting and admissions at about a 40,000-foot level. We haven’t gone down into detail, and that’s okay. We’ve talked about student support and the IT integrating the systems across campus.
[00:20:22] Alumni and Donor Stewardship AI
Drumm McNaughton: The other one that I thought was fascinating was your alumni relations and donor stewardship.
Let’s talk about that a little bit.
Justin Beck: Yes, this is a area that has been a little slower in the uptick of a-adopting AI technologies to support their processes, their systems, and their intelligence. Some of it has to do with legacy systems, quasi-old ways of doing things. But there’s leading institutions like, Boise State University, for example, that has incorporated numerous AI tools to help with their fundraising efforts.
Example, one of the best examples is they provide to their major gift officers an intelligence tool that helps them target, within their portfolio, when is the right time to be engaging with. Major gift officers have roughly a hundred fifty to two hundred people on a portfolio that they may work with, and th-that’s a large number of prospects, if you will, to constantly be engaging upon.
So everything from creating custom emails that should be sent on a regularity, insights as to what’s going on with all a hundred fifty to two hundred of those families or individuals or family offices. Boise State was one example that both from targeting around major gift officers as well as h-helping curate earlier in lifespan giving experiences, saw an eighty si-seven percent increase in donors. The volume of donors year over year, with a fifty percent increase in the number of donors that have a repeat giving experience to the institution.
That generated multiples of millions of dollars to the organization. And there’s a lot, obviously, Drumm, as you’d expect, that went into that, from curating personalized websites, individual personalized outreach that was fueled by AI, as well as a wealth of analytic knowledge being surfaced to those that interface with the donors.
So, really cool experience.
Drumm McNaughton: It’s amazing that one thing you really caught me, because I’ve got a good friend of mine does fundraising and coaching people on fundraising, is when is the right time to approach a prospective donor? And if you’ve got AI who can put you on a calendar, develop a calendar, and give you the things that person is interested in, I would think that would increase your likelihood to receive a donation, especially a sizable donation.
The sizable is really about building the relationship, but you’re using AI tools to start and build that relationship.
Justin Beck: Good Drumm, exactly. The mantra in the fundraising space is honestly your first interaction with a potential donor shouldn’t be an ask. It shouldn’t be an email asking for their donation. The concept of like building stewardship over a time period with them, whether you engaged an alumni, a family member, a parent in volunteer or mentorship opportunity, a athletic participation opportunity.
There’s numerous ways through events, in either campus or virtual experiences, where you s-start to cultivate stewardship and affinity to the university. And one of the, one of the big trends we’re seeing in our alumni and fundraising world is actually, the tools helping engage students before their graduation date.
So getting them involved in mentoring opportunities, leverage, having the alumni department sponsor, mentor matching, which AI has been a huge boost for us to help with like matching alumni to students. So, they build this affinity that, gosh, my university helped me with my employability, my mentors that I work with, whether I’m a current student or I’m a graduate looking for– a young graduate looking for a new mentor.
AI tools are, are– have been thrust into the center of that, of helping take these hundreds of thousands of alumni and tens of thousands of students and helping with the interest matching and ability matching. So that can… And I know that discussion is so far away from like fundraising, but those are the type of activities that make someone’s propensity to give back to the, back to their alma mater, increase by hundreds of percent, when they’ve had those type of experiences.
Drumm McNaughton: Yes. Well, and especially because fundraising is a long-term game. It’s not, like you said, you don’t do an ask on the first time around. It surprises me that when I speak to so many prospective clients, the first time they say, “Well, give me a proposal.” It’s like, I don’t know you well enough to give you a proposal at this point.
But it’s, anyway, we’ll get there. There’s some gotchas with this.
[00:25:15] Security Policies and Guardrails
Drumm McNaughton: And I would think one of the big ones has to do with security
Justin Beck: Yes. If an institution hasn’t established very clear guardrails from a policy perspective about acceptable use, the type of systems, that, that are available, like you’re seeing just Google search, like higher education, AI acceptable use policy, and you will see and, yes, I would put well, well more than fifty percent of institutions have something in place, and the rest of them are working on something right now.
So custom critical that, that is like documented both from a liability perspective for the institution, but as well as for guidance to the community, how to do it. The second part of this, Drumm, that I think is like also so, so incredibly important is to take the bull by the horns. Don’t let in the institution, or organizations throughout the institution run wild with picking up all their own individual things.
There are very clear areas, and we’ve talked about some of them today, where the institution itself can sponsor the model, the tools, and some of the centralization of resources that help keep their very private and secure data captioned within the organization. Mm-hmm. Exactly. Exactly.
Exactly. Great example
Drumm McNaughton: One thing that I am seeing, and I’d like your opinion on this, is that many institutions are getting their own private instance of these tools so that they aren’t sharing their proprietary information across multiple platforms. Are you seeing this as well?
Justin Beck: We are. It’s a way to put up some of those guardrails to not only protect the institution, but also protect the students, and faculty from those type of pieces. And you’re able to… Some of that curation that we talked about before, when you’re not as exposed to, “the outside world,” it’s giving, like, a really strong boost.
Now, I think you and I both know that’s probably happening at some of the tier one top Carnegie classifications of institutions compared to the larger piece, because sometimes they have more assets at their disposal, but it’s absolutely a trend we’re seeing across the board. And I’m very intrigued to see how that evolves as the public facing pieces continue to evolve.
Drumm McNaughton: Mm-hmm. I am as well.
[00:27:34] Human in the Loop Implementation
Drumm McNaughton: One of the things that we’ve talked about is the importance of requiring a human be in the loop for the decision-making, not turning it over to AI completely, but you’ve got that human ethical input as well as people making the final decision saying, “Yes, this is a good idea,” or, “No, this is ridiculous.”
Justin Beck: Correct. Correct. I’ll harken back to the example I gave of like, again, very tangible example, a virtual assistant chatbot that’s sitting in support of a operational unit inside a university like financial aid or recruiting and admissions. There isn’t a set it and forget it capability where you turn on something and it auto-magically does all the work that you’re hoping it to accomplish for like after-hours support.
No Staples Easy Button here.
Drumm McNaughton: Auto-magically?
Justin Beck: Auto magically.
Drumm McNaughton: I like that. I’m going to borrow that one.
Justin Beck: Take it. Oftentimes when we’re interfacing with prospective customers who are like, “I just want to put this thing in place and have it do all the work,” I’m like, “It’s still going to be the Wild West if you, if you let it just jump in and do all the work.”
Human intervention there to make sure, like I mentioned, about reviewing things like what are the most– 20 most common questions that are coming into my department? When are they coming in? What was the next action that student took after that question was provoked? Those are all things that half an hour a week, someone just sitting down reviewing that and updating some of those workflows or the guidance or the knowledge of content.
Drumm, I got one… I know I’m going long on this answer, but I want to share one beautiful example of where human intervention comes into play.
Universities are littered with websites and properties and often and as you can imagine, oftentimes those have not been updated to the most recent information.
We’re working with a school in Missouri most recently, and we put our web crawler in place that goes out and looks for all their institutional content knowledge, and they found that there were three different dates published on live active websites about when recruiting and admission deposits were due.
Three different dates. So, think about that from a student perspective. So again, human intervention. Boy, we took down two. We updated one of those web pages. Solved. But they wouldn’t have found that out until deposits were late or deposits were coming in at the wrong time. And it sort of makes the institution look foolish when they’re not aligned. Like simple AI tools crawled, exposed that information, and allowed the organization to take action.
Drumm McNaughton: And each week, each month, whatever, AI can give you a list of where people have gotten stuck so that you can go in, it can make recommendations for you as to what needs to be fixed, but it’s still that human that needs to be in the loop, and that’s one of those critical things for implementation.
Justin Beck: Absolutely. Drumm, think about the continued changes in financial aid policy, legislation, things that are coming at a federal level. Every year, there are significant changes. How does an institution keep up with all those requirements, those change of expectation, the change of classifications?
AI surfaces all that, updates knowledge repositories, updates answers, can update web pages. It’s savings two FTEs full-time jobs to keep up to date on this and allow the institution to get focused on those higher impact personalized touches that they want to provide to students.
Drumm McNaughton: And that is critical. Enrollment, fundraising, whatever it is, the student services, it’s got to be that personal one-on-one, but there needs to be a way for people to… Everybody’s busy. There’s got to be a way for AI to guide the person with the question to the right answer, and if that doesn’t work, then that’s the time for a human to get involved.
Justin Beck: Absolutely. Absolutely, Drumm. I think, this comes to the implementation discussion we had. I think where institutions have probably faltered at times with AI-based technologies is that they haven’t had an end goal in mind most of the time. It was the new whiz-bang toy that is available to be as players.
Like any good project, you’ve ran many a projects over your career, it’s start with the outcomes in mind, and then we’re going to design systems and tools around how to make them most successful. So, that’s one of the key pieces of advice we start every one of our engagements with.
Drumm McNaughton: That’s great. Well, Justin, this has been a fascinating conversation for me. I apologize for throwing a curveball right there at the beginning. That’s a question. The AI bots just drive me crazy sometimes.
Justin Beck: You’re one of the hit zero… I do it too. When I call in, I just start hitting zero sometimes just to get me out of question and response.
Drumm McNaughton: Yes. I find that if you get through that three times and they haven’t given you to a live body, it’s not worth it anyway. I’m just going to, I’m just going to never give that product a recommendation.
Justin Beck: I love it. I love it. I love it.
[00:32:41] Three Takeaways for Higher Education Presidents and Boards
Drumm McNaughton: So as we draw down to a close our two takeaway questions, three takeaways for presidents and boards, what do they need to be thinking about from an administrative perspective and the AI tools that are out there?
Justin Beck: Absolutely. One, move quickly, but with the things we talked about. A policy in place, a thoughtful design of your outcomes that you’re ultimately trying to drive to, and these are going to come back to your common institutional outcomes of student success, enrollment success, alumni success. So focus on those outcomes, but move quickly to not let things sprawl and get out of control where you’re having to wrangle things back in.
Two, control what you can control, but keep your stakeholders in mind. Like, pull them in. This is one of the most beautiful opportunities for… it used to be my least favorite term in the world, faculty committee, and now I love faculty committees, because sometimes they can get a little unruly and a little, like, pull them in for both the spectrum of neophytes all the way to the experienced.
And then I think the last piece of advice I have with folks is iterate often, is that the technology is changing on a month-by-month basis. So set up, like, a governance structure that says, “We’re going to come back together at the end of June to see how the effectiveness of these programs are running and what pivots do we need to make so we keep those institutional outcomes in mind.”
Drumm McNaughton: That’s really, all three of those really excellent. Especially get started and get started now because you’re– if you’re thinking about waiting until the technology slows down, forget it. It’s not going to slow down for at least five to 10 years.
Justin Beck: Drumm, the comparison from the athletics world I think about is like, think about like, I don’t want to throw shade on something, but think about that the NCAA took a long time to get their hands around this NIL situation, in, in, in higher education right now, and that has spawned everyone solving it for themselves, which means no one’s solved it right now.
So opportunity for institutions to get their hands around this early.
Drumm McNaughton: Yes.
Exactly.
Absolutely. What’s next for you? What’s next for Gravyty?
Justin Beck: Oh gosh, we are, continuing to infuse more AI-based technologies into all the products in, in our portfolio. Doing a lot of work with integrations with some of our key partners like Ellucian and Salesforce and Blackbaud that work across that full student life cycle. What’s most important for me right now is to be running around this great country and this great world, having a lot of conversations with institutions that are focused on improving student success and engaging with a ever disconnected alumni population.
And so trying to come up with what are the modern day strategies that are going to help us really engage with the constituents that we all have the honor that we get to serve.
Drumm McNaughton: Yes. Well, Justin, thank you. I have thoroughly enjoyed this conversation. I certainly hope our paths continue to cross in the future
Justin Beck: I hope so too. We’ll have to find our way to Annapolis at some point in time and, spend some good time there, so.
Drumm McNaughton: Yes, sir.
Justin Beck: Awesome. Drumm, thank you.
Drumm McNaughton: Thank you. Take care.
Thanks for listening today, and a special thank you to my guest, Justin Beck. Justin, thanks for joining me on the show. I thoroughly enjoyed our conversation. I look forward to the next time our paths cross. To my listeners, thanks again for tuning in. See you next week.
