Beyond Traditional Models: New Signals for Driving Student Connection and Yield

Getting a signal is easy. Knowing which ones actually mean something is the hard part.

That was the theme for EdSights' recent virtual panel, Beyond Traditional Models: New Signals for Driving Student Connection and Yield — a conversation that brought together three enrollment leaders and EdSights' co-founder. The panel examined what's really predicting student intent today, and what changes once student voice becomes part of that picture.

Panelists:

Dr. Paul Orscheln, Vice President of Enrollment Services, Park University 

Justin May, Chief Enrollment Management Officer, Richard Bland College 

Tom Nesbitt, Vice President for Enrollment Management, Clinton Community College 

Carolina Recchi, Co-Founder and CEO, EdSights

‍

The signals that matter most come with context

Click data and email opens have been around for years. What's new is how much earlier teams can catch a student pulling away, and how much more context they have when it happens. Justin May's team tracks everything from stage velocity outliers to a financial aid email opened three times with no follow-through, treating each as a piece of a larger pattern rather than a standalone alarm.

"The tools we now have at our fingertips through AI-powered predictive models, things like that, allow us to get to the why, so we can have those more personal communications and hopefully intervene before disengagement becomes permanent, which I think is the why."

Justin May added that no single signal tells the whole story on its own. It's when several show up together — a stalled application, a drop in engagement, a question left unanswered — that his team knows it's time to step in.

‍

Students are more honest than institutions expect

Ask enrollment teams what surprises them most about listening to students at scale, and the answer is consistent: students say things they'd never bring up on their own, and the concerns rarely match what the institution assumed.

"These were model students, high GPA, good attendance, work-study students in enrollment management. From the surface, it seemed like nothing's wrong. They're going to float by, they're going to graduate, they're going to move on. But after speaking with them, we identified that the events designed for students felt like they were created for our residential students, and not the commuter students, so it was creating that disengagement, and we were at risk of losing them.

Justin pointed to a bigger reason students open up in the first place: a passive way to raise a concern removes the fear of judgment that keeps a lot of eighteen and nineteen-year-olds from saying what's really going on to staff or faculty directly.

‍

Alignment matters as much as the tools themselves

Knowing what students are really dealing with only helps if the people who can act on it are actually talking to each other. Every panelist agreed on one thing: new technology doesn't solve old problems on its own. It just makes whatever process is already in place move faster, for better or worse.

"At a place like Clinton, our care team is an office of one or two, so it's fairly easy if you're reporting to the same VP, to get those care teams together to walk through that process and create the opportunity for the staff to participate in that change, to make the big decisions that are going to help the students longer term. When we're looking at it, we're having weekly meetings, making sure we're working through each process through the course of the year. The focus may shift through the course of the year from one office to the other. But all hands on deck. Many hands are going to make lighter work throughout the process."

Paul Orscheln has seen the same pattern at Park University, where financial aid, student life, and admissions all need visibility into which students are struggling, since students rarely show up with just one problem.

‍

Figure out your ideal process, then let AI scale it

That alignment only pays off, though, if the process everyone's rallying around is actually sound. With so much noise around AI in higher ed, Carolina pushed the panel to separate what's actually working from what's still more promise than practice. Justin pointed to real gains already showing up: identifying signals at scale that a single counselor could never track across a full caseload, and lifting enough administrative load off staff that they can spend more time on the conversations that actually move students forward.

"AI can help you scale process, initiatives, and workflows, but if those processes are broken, or there's no best practice for it or there's no guidelines for it, either you're not going to be able to scale them because the AI does need some directions and guidance, or it could scale something that's broken. Figuring out, 'what is our process or ideal flow?' and then plugging AI into it, as opposed to hoping the AI will do the initial part, has been key for our partners."

Justin agreed, and was direct about where the limits are: AI isn't a relationship builder. That part still has to come from a person, with the technology there to support it.

‍

What gives these leaders a clear next step

Asked for one concrete change enrollment leaders could make in the next 90 days, the answer that stuck: add a single concern or motivation question to your application. It costs nothing, and it gives your team the context to understand what a student actually needs before they ever ask for help.

That's the throughline of this conversation. The signals are already there. What separates institutions seeing real results isn't the sophistication of their tools, it's whether they're listening for the right things, believing what students say, and building the alignment to act on it.

Want to see what Student Voice Intelligence could do for your institution? Click here to learn more.

‍