Jeff (4:17)

Yeah, I agree. One of the biggest moves, I think, in the market lately was the combining of Ivy, AI and Ocelot under Gravity. People on the outside see a merger, people on the inside probably see strategy. What was the vision behind bringing these chatbot and AI platforms together?

Sharon Harrison (4:35)

Yeah, I can't take credit for the vision. I can speak to it because it has been a journey and transformation over the course of this year. So as a little bit of background, Ivy and Ocelot, I would say, were the number one and two. Depends on which company you ask. Virtual assistant providers specifically purpose built for higher education. And Gravity is the leader in alumni engagement and fundraising. So when you look at them disparately, it's like, well, that doesn't make sense. But when you really think about the student life cycle, the higher education journey from prospect through to engaged giving alum, then it makes sense, right? How are we building the community from the second we first engage with the prospect all the way through to driving institutional loyalty, you know, with our alumni and their experiences.

Jeff (5:31)

I want to ask you about this evolution. When chatbots first hit higher ed, I think a lot of people treated them like digital interns or give them a list of FAQs, point them at a website, hope they handle the basics without breaking anything. But over the last few years, it's really been a turning point. Students are expecting immediate help, campuses are juggling record complexity, and AI has really moved from simple intent matching to true conversational capability. So it feels like we've crossed from nice to have territory to core infrastructure, especially for enrollment and student services from where you sit, how have you seen the role of chatbots evolve in higher ed over the past few years?

Sharon Harrison (6:15)

Yeah. Well, I'm gonna, I'm just gonna ask you a quick question for my own data collection purposes. Can you recall or make your best guess about the first chat bot that you interacted with in any business sector? Probably maybe like your bank or your cable.

Jeff (6:31)

I think it was retail. Like maybe, I think on a Dell or Lenovo website or something like that is what I would think some high purchase I was making. There's a, like help. Ask me a question. I think it was something like that.

Sharon Harrison (6:41)

Got it. Okay. So that's pretty, I think, the sort of baseline experience, right? It's some kind of customer service. I mentioned your bank, your cell phone provider, a retailer. Right. And so that was the preconceived notion. And so I recall fondly, you know, that the first calls I was on were, well, what is a chat box? And it was a box, not a bot. And I would think, okay, I understand now we're going to need to do some education about what is possible. And so it became really important to kind of understand where the conversation was starting from. Right. This is six years ago now, and AI was a novelty, or like you said, it was a. I mean, maybe it's not even a nice to have at this point. It was like a little bit of an unknown. And now it's part of the zeitgeist. It is table stakes. You know, I think there were a few things that happened that really drove the quick evolution because it does feel pretty rapid to me. One is, you know, the enrollment cliff, summer melt, all of these like massive factors that are directly impacting institutions and also students. Right. Like, I never, never crossed my mind that institutions would close, would close their doors, shut down campus. But that is the world that we are living in. And I think we started to see the like, indicators of that years ago. Covid. Covid drove really rapid innovation. And so it forced institutions to become digital campuses and to drive that online experience. And so that happened very quickly once we were a few months in and it was like, well, this isn't going away. We've got to figure something out. Adoption became part of the ethos. And then I think we saw open source, LLMs, GPT. And that was the boom. That was when there was an interest in adoption, but also, of course, a heavy amount of caution. So we went back into the education phase again. And that's where we saw the evolution from a chat Box to a chatbot to a virtual assistant. And I think now, and I'd love to hear your thoughts and what you're hearing, I think now we're seeing the shift from a virtual assistant to an agent as we head towards agentic and that like fully autonomous experience. What are you hearing? I know that you really are part of a lot of really good conversations.

Jeff (9:16)

Yeah, I love Chatbox. I hadn't heard that before. So, like, that's funny. That was, that was how it started. But one thing I've noticed talking with campuses that I think is worth mentioning is that pain points aren't the same everywhere.

Sharon Harrison (9:28)

Yes, great point.

Jeff (9:29)

Some schools are buried in enrollment inquiries and missing prospects because staff can't keep up. Others are watching retention soften and need more consistent outreach to the students who are slipping through the cracks. There are advising offices that are stretched thin trying to support students who need answers at, at all hours. And across all that, I think AI is touching more parts of the student journey than ever. No one's arguing that, but it's no longer just the front door support. It's becoming embedded across the whole life cycle. So I guess I'll throw it back at you. What are you seeing? Which use cases are really gaining traction right now, Whether it's enrollment or retention or advising or something else entirely, that's.

Sharon Harrison (10:10)

What makes Gravity's world complicated and fascinating, I think, because there are really, just like you outlined, there are use cases all across campus. And those are the things that keep administrators up at night. It's what restricts their access to funds. Certainly this year we've seen a tighter correlation between retention outcomes and how much funding you're able to access. So at a higher level, it's about staffing shortages and budgetary and operational efficiency. But then we do get into the very department specific use cases, right? They're all dealing with some sort of budget insecurity, I'll call it, and some sort of staff shortage in some capacity. But it is, you know, Admissions is really thinking about how do I increase the application submission rate. Right. If you work with an online institution, it's about, we know that we have a higher rate of submission if a student starts and completes their application within 24 hours. Well, you definitely don't have 24 hour human support. So what do you do in lieu of that financial aid? A completed financial aid application is the highest indicator of enrollment. So how can we power those things to deliver against the KPIs that the institution is thinking about? Gravity's in this great space of like B2B2C. Because we are partnering with the institution, those users have specific needs. But also we're thinking about what the student, the customer needs.

Sponsor Announcer (11:54)

And now a word from our sponsor.

Sponsor Voice (11:59)

How can your next campaign soar? With experience helping colleges and universities raise billions of dollars, Mackie Strategies delivers communications, fundraising, and tech expertise that your campaign can take to the bank. Mackie Strategies, build your breakthrough.

Jeff (12:21)

I think what's striking about the schools you work with is how much ground that these AI tools cover. They're not just answering questions, they're catching things staff might miss. Supporting students who study late at night, helping families who speak languages the campus team may not. So when you add proactive outreach on top of that, I think you start.

Narrator (12:45)

To see moments where technology actually changes a student's trajectory.

Jeff (12:49)

Those stories are often the ones that stick with people. What's been the most surprising, impactful result you've seen from providing 24. 7 multilingual support or outreach?

Sharon Harrison (13:00)

Yeah, I'll give you two examples, if that's okay. One is just a little bit more anecdotal. I was in a conversation several years ago and we were talking through use cases, and this particular team was tasked with, in part, addressing food insecurities on campus, housing insecurities. How can we get information to those students to help reduce that obstacle for them, to allow them to focus on their schoolwork? I hadn't even thought about that as a use case. But when you think about the shame element or the fear of judgment and not wanting to ask if you can talk to something that has the information and is available when you're available and is not going to bring shame, it's going to be very empathetic. What an interesting way and very humanist way to deploy technology. So that's kind of just a little anecdote. You know, we know that on average, 32% of calls come into two financial aid teams outside of business hours. So let's think about what that means if. Let's look at the holidays next week, right. I hope that everybody is out of office and enjoying some downtime with family, friends, whatever. But that means they're not answering the phone and answering emails. And this is a time of year when I know it is in my house, when we are making sure our FAFSA information is completed. Right. Or applications are being submitted, all these things. And so if I have a question that is preventing me as the parent from completing that paperwork, well, now I have to wait until you're back in the office, but I have limited time to attend to that. And I work full time, so I can only do it in the evening when no one's available. And so now it puts my daughter's application at risk. And I, I have a very fortunate experience. Right. Like I'm a college grad. I do have a full time job. I'm not working a day and a night job. I speak English. Right. I have all the things going in the right direction for me. But if 32% of your prospective students are just waiting to find out an answer, that's a lot.

Jeff (15:16)

Yeah, that's powerful. That's, that's a. Interesting perspective. You know, one theme I hear non stop from digital leaders is that they're.

Narrator (15:28)

Already running a fragile ecosystem.

Jeff (15:31)

They've got a CRM that talks to their sis, an advising system that sort of talks to their lms, maybe a dozen other tools stitched together with duct tape and optimism. So when a new AI solution shows up, the first reaction is often, please don't break anything. Schools aren't just buying features. They want confidence that this won't add friction to the systems that they already feel maxed out on. So with that in mind, how do you make sure your AI tools fit cleanly into what campuses already use, whether that's their CRM or SIS or other core system?

Sharon Harrison (16:08)

Jeff, I love that because what you're highlighting is the sort of fragility and the nuance of a higher ed digital ecosystem. Right. Like, I think we often picture higher ed as the. Well, at least I do as this very sort of sacrosanct thing. It's a business. And these are the systems that are being used by these individual business offices because they are very focused on their goals and outcomes. I'm curious, I wonder if you're hearing the same things in the market that I am, where there are campuses that are like, we had no idea. We have five different CRMs across campus because our grad school is using a different one than our medical school, than our. Right. Like, have you.

Jeff (16:51)

Oh, yeah, yeah, that's how it is. And the old I think answer that was like, oh, yeah, we can consolidate. And I think a lot of schools have given up. They're like, it's probably not going to happen. I mean, they can try here and there, but for the most part we need to kind of accept it. Yeah, consolidate where we can. But yeah, I'm seeing.

Sharon Harrison (17:06)

Yeah. And right. Like the concept of consolidation, I don't even want to think about trying to undertake that. So what we try to do is sit on top of all of that. Right. It's about taking what's already there, all the time and effort and care that has gone into building those systems and then feeding the data and ensuring that there's some sort of validity and hygiene around the data. We don't want to disrupt that to your original question. We want to capitalize on that and make that data worth even more to the institution and layer some insights on top of it. So it's about plugging in and making it. When I'm trying to get in for. Not me, I wouldn't be allowed to, but when my daughter is trying to get information about her application status, what if there's an integration between the CRM and the virtual assistant? And so now I can say, what's the status of my application? I can be asked to authenticate through sso. And now the bot, the chat box can come back and tell me it's pending review or whatever it might be. Right. So that's the objective, is to take the CRM, sis, erp, lms, itsm, housing system, HR systems. Right. To take all of those and go. We're not here to replace them. What we specialize in is sitting on top of them. And we are not use case or department specific in a good way. We can get that granular and be specific to it, but we know how to work with all of. We have experience working across all departments. That's a really important thing to think about, because the reason institutions have landed with five CRMs is because nobody's talking to anybody else.

Jeff (18:51)

Right.

Sharon Harrison (18:52)

And so we're trying to bridge that gap.

Jeff (18:54)

Yeah. Digital governance is, is an ongoing challenge in the silos. And, you know, a lot of campuses still measure chatbot success the old way. I think they'll say it answered 12,000 questions this semester, which sounds impressive until you realize it doesn't tell you who got helped, who gave up, or whether it actually changed anything about student behavior. So the schools that really get value out of AI tend to look deeper. They look at where the bot prevented a case from being created, where it nudges student to take action, or where it uncovered gaps in content policy that staff didn't know were there.

Narrator (19:33)

Maybe.

Jeff (19:34)

And those are the signals, I think, show whether the bot is helping the institution move forward, not just fielding volume. So I guess my question is, what metrics should institutions focus on to understand true chatbot effectiveness beyond simple question counts?

Sharon Harrison (19:51)

I love the metrics that you bring up because you're connecting the dots, right? You're thinking about the whole B2, B2C experience. You know, we do ask our partners to pull some benchmark or baseline data. You can't know if you're making progress if you don't even know where you started. So how many calls are you getting? How many hours are your staff spending? Responding to emails? How many minutes does it take for them to resolve an issue with a student? Those kinds of things become really helpful. Baseline data. Houston recently deployed for four of their 11 departments. We're working, you know, we've got to do it in phases. And already between February and September, so however many months that is six, seven months, 32,000 calls were deflected. Okay, so that's a number. But what does it truly mean when we work it through? You know, we've got to start thinking about this in terms of dollars and cents, Right. How does that impact their bottom line? Yes, it frees staff up to have higher value conversations. Right. If we can deflect the tier 1 and some of the tier 2, then staff should be spending their time on the tier three. Questions increase problems. But if we roll that 32,000 out and look at 3,000 or so hours saved and then we take some really conservative salary estimates, there's about $50,000 of savings there just from four departments, just from a very simple rollout. What's driving that, though? And this is a little bit of an aside to the metrics point. You also have to be thinking about how are you driving your students to use it? Right. It's the old if you build it, they will come thought. Right? Like, you can't just stick it on the website and go, oh, we really hope someone comes and uses it. You have to push the adoption. And so schools like Houston or Lone Star ran social media campaigns for their students. They spun up commercials. So we expect to see that type of engagement, that 32,000 deflection point. But then we need to look at things like how much time would it typically take to onboard a student worker that's going to be answering calls? Okay, well, what if we could onboard them quicker and get that time to value shortened? You know, how many questions do your back office staff members field about benefits and do we have people doing their open enrollment in time? Right. So all of those things that really comes down to like, what is it that your boss is going to be asking you about? Let's tie it to those data points. Right. What's keeping you up at night? So it's very, very customized, but on a high level, it's always about how many chats will it get how many unique visitors? Oklahoma admissions used it years ago to kind of distill out demographics so they could understand. Okay, most of our engagement comes from these areas. I'm making this up. I don't remember, truthfully. You know, these areas outside of Norman, we should run special recruitment events at high schools in that area to drive increased enrollment because we know that there's interest there. So there's all different ways to pull the data out and analyze it again. It's just tying into what, what matters for your team.

Jeff (23:17)

Yeah, I think when people hear AI chatbot, I often picture a single system doing everything on its own. But in reality, campuses need something more flexible. Some questions are perfect for automation, some.

Narrator (23:31)

Need a human touch, many are in the middle somewhere.

Jeff (23:35)

And I think that's where hybrid AI comes into play. It's not about replacing staff necessarily. It's about routing each interaction to the right level support so students can get faster answers. They can get humans in the loop when they should. The best systems, I think, make that handoff invisible to the students. So can you break down what hybrid AI really means in this context and why it matters for student engagement?

Sharon Harrison (24:02)

Sure, yeah. That invisibility is, I think, important. Right. Like you don't want it to feel disjointed, you want it to feel very seamless because that builds confidence and trust in the system. So for us, that hybrid algorithm is really taking what is retrieval based AI. So earlier you mentioned like that FAQ kind of action. Right. It needs to be very static. Ask what's the FAFSA code and you would get back the code. Right. That was retrieval based. It was all kind of prescribed in the bot's knowledge. The hybrid element is the result of generative AI being available. And so that hybrid motion is, you know, the way we build our technology is we utilize the institution's knowledge sources. So you've already gone to the trouble of building up thousands of pages on your website. Let's use that. That's your source of truth. And we'll use that institutional knowledge to power your bots responses. So where, if you go to GPT or Gemini, the knowledge sources powering those generative algorithms are the entire Internet. And all of the information people are putting into those models, that doesn't work in higher ed. That opens a lot of risk. And so the guardrails around our technology are part of this algorithm. We only use the institution's content. So the institution is the one saying, yep, this is the prescription to solve that problem. And the bot won't go Outside of that and the flexibility of like curating the responses that are really important and need that tight curation, we can pull those in with the retrieval responses. So it's a handoff between the two, but very contained and allows for that conversational feeling. It's invisible, right. You don't know what the response is coming from, but it's conversational, can answer complex questions. So that's another aspect where it's not, you know, you can only ask one question at a time. That's not really how humans talk. Right. You ask questions like how do I apply? And what's the application deadline? Well, that's a complex question. And so we need to respond to that in a conversational way. The last point I'll add is that part of the algorithm comes down to the tone. Institutions take a lot of care and time to define as a marketer, you know, that's brand guidelines and how you want the institution to appear. That goes into how they write everything and how they would want their bot to respond. So we can also curate the bots tone. If you want to sound like a pirate, we can do that. Right? We're open to it.

Jeff (26:52)

Well, you know, anytime AI shows up in higher ed, I think the conversation quickly moves past features and into responsibilities. Leaders want the efficiency, but they are also thinking about FERPA and ADA and equity gaps and the reality that AI can sometimes hallucinate or reinforce bias if it's not handled correctly. Schools can't afford to make real, I mean, mistakes here. I think they need the confidence that the technology is safe, transparent, aligned from policy from day one. What guardrails do you recommend schools put in place to address data privacy, accessibility and bias as they adopt AI.

Sharon Harrison (27:33)

Yeah, so you kind of called out the big ones, right? You're going to want to make sure that you look at the vpat, got to be current and I would say have it be a bidirectional conversation. It should be collaborative. You as the user are on the front line. So if you see something that is not accessible, you should be able to go back to your vendor and say like, hey, we're helping you. This is not meeting guidelines. Ferpa hugely important demos are a great way of kind of sussing out, you know, the FERPA knowledge of the product team. A SOC2 certificate is really helpful these days. Right. I'm sure the security and technical review teams appreciate when you can come through the door with that look more than anything. I think it's about ask a lot of questions, make sure you Understand, what is the algorithm? Where is it pulling information from? How is it being delivered? I love that you even thought to ask about our hybrid AI because it shows that you are paying attention to what does that algorithm look like. That's, I think, probably the question you should lead with. Right? Like, what's powering your technology? Where does the AI actually show up? Because that's the other thing. All AI isn't created equal. And just because you say you're using it doesn't tell the customer how you're using it.

Jeff (28:55)

Yeah, you know, every year a new capability appears. Voice AI, generative tools, predictive modeling. It feels like we're entering the new choose your own adventure era of chatbots. What's the next frontier you see coming for higher ed AI?

Sharon Harrison (29:12)

Okay, choose your own adventure books, Arguably the best books of childhood. The future, I think we're already seeing a little bit of it, but I don't think it's as good as it will be. It is to your earlier point, about institutions wanting to identify those students that are at risk. Well, there's decentralized data all across campus. And to really understand if a student is at risk, shy of asking them and hoping that they tell you a truthful answer, you've got to be able to look at the full student profile. Right. You want to be able to see, are they late paying their student account balance? Have they been going to class? What's their gpa? Do they have a full meal plan? Are they first gen student? Let's look at across their entire, you know, who they are instead of just looking at, you know, they're getting straight A's. Okay, yes, that's great. But that doesn't always guarantee an outcome. And so if we can be paying attention to all of the signals students are giving us and anticipate and initiate. So if I would love it if my daughter, who's a junior, was not going to class and her GPA was dropping around a 2. And you know, if she got a notification, if she got a text message because she doesn't pay attention to anything else, she got a text message and said, hey, you know, we have tutoring services available at this time in this building. They're free. Please come. We would love to help you. Or don't forget your professor has office hours at this time. Like, let them know how much you care. That is what these students want. They want to know that there's mental health services available if they need them, because they might not be willing to ask. So I think that's the thing. Let's take away the assumption that they're going to advocate for themselves and let's intervene earlier. Goes back to the whole, why did I want to do this? It's about democratizing access. If you don't know it's available, you can't take advantage of it.

Jeff (31:15)

I love that. Well, to wrap it up, one last question for you. Not every institution has a huge team or a massive tech budget. Some feel like they can't even get to step one. For a smaller institution with limited resources, where would you recommend they start? With a chatbot or AI adoption?

Sharon Harrison (31:32)

Look, I mean, the salesperson in me wants to say, give me a call. We'll figure it out. But in fairness, look at what you already have. Right? Like, even I'm guilty of this, right? Our marketing team uses different things. I don't even know what's fully available until I think about, like, well, gosh, we were going to stand that up. Is that already available in HubSpot? I'm not sure. Do that audit. See what other areas of campus are using. I know that that is difficult, but getting as much information as you can for what's already in place will ultimately shorten the procurement cycle for you. I mean, gosh, if you could avoid going to rfp, wouldn't you love that? Like, I don't think that's on anybody's go card for. Well, I shouldn't say that. I'm sure it's on some, but that would be my suggestion because then it's also gone through the security and technical review and you know, you have some shortcomings to take advantage of.

Jeff (32:24)

Yeah, that's good advice. Well, thank you. That was the highlight of my week, too. Sharon, great to have you. I will put links to Sharon's LinkedIn as well as Gravity's website in the show notes. So, again, great talking to you, Sharon.

Sharon Harrison (32:38)

Yeah, thank you, Jeff.

Sponsor Announcer (32:41)

We wrap up this episode. Remember, EdTech Connect is your trusted companion on your journey to enhance ed education through technology. Whether you're looking to spark student engagement, refine edtech implementation strategies, or stay ahead of the curve in emerging technologies, EdTech Connect brings you the insights you need. Be sure to subscribe on your favorite podcast platform, so you never miss an.

Jeff (33:04)

Inspiring and informative episode.

Sponsor Announcer (33:06)

And while you're there, please leave us a review. Your feedback fuels us to keep bringing you valuable content. For even more resources and connections, head over to edtechconnect.com your hub for edtech reviews, trends and solutions. Until next time, thanks for tuning in.