Campus Technology Insider Podcast December 2024
Listen: How Generative AI Is Answering Student Questions at Bryant University
Rhea Kelly 00:00
Hello and welcome to the Campus Technology Insider podcast. I'm Rhea Kelly, editor in chief of Campus Technology, and your host.
At the start of each academic year, a thousand new Bryant University students come to campus brimming with questions about everything from class registration and building locations to dining hall hours and WiFi connectivity. And thanks to the power of generative AI, they can get their answers from Ask Tupper, a next-generation chatbot designed from the ground up at the university. For this episode of the podcast, we spoke with Chris Stephenson, managing director of intelligent automation, AI, and digital services at AI solution provider alliantDigital, and Chuck LoCurto, VP for information services and CIO at Bryant, about how Ask Tupper started, what's possible now with AI-powered chatbots, lessons learned from the project, and more. Here's our chat.
Chuck and Chris, welcome to the podcast.
Chris Stephenson 01:13
Thanks for having us today.
Rhea Kelly 01:16
So I thought we'd start by just having you each introduce yourself and tell me a little bit about your background.
Chuck LoCurto 01:23
Sure, I'll go first. This is Chuck LoCurto, I'm the Vice President and Chief Information Officer here at Bryant. I am coming up on 13 years. I spent 20 years at Textron, primarily Textron Financial and Textron corporate enterprise apps. I was, sort of my 20 years there, I spent seven, my last seven there as the CIO for the finance segment. So I've had the privilege of having this, this IT role for the past, what is it, I guess, nearly, my gosh, nearly 20 years. Kind of crazy. I do a lot of other fun things, fun things at Bryant that usually catches most people's attention. I coach the Division 1 diving team as part of the swimming and diving program. I started and coached it for eight years, and now I just help on occasion. But my background, you know, my background is in IT. Got my masters at RPI back in the day. And I think that probably summarizes enough. If you have more questions, feel free to ask. I can go deeper.
Rhea Kelly 02:29
Go deeper with the diving background, please. I can't resist. So Chris, how about you?
Chris Stephenson 02:38
My name is Chris Stevenson. I, I'm the Managing Director here at Alliant on emerging technology, which includes AI, automation and digital services. I've spent really the last 25 years of my career in emerging technology, ranging all the way from when the internet first came out, through the, through the mobile phase, through the social phase, and now through this, this new AI phase. I work with most industries with this type of work, but I always find myself, especially as new technologies really starting to emerge, working back with with different higher ed organizations, because that's where, honestly, a lot of the technology starts to become real.
Rhea Kelly 03:18
So earlier this year, Bryant announced the launch of a new gen AI chatbot called Ask Tupper. So could you just kind of take me back? How did that project start?
Chuck LoCurto 03:30
So, you know, that's interesting. Sometimes these things are a blur, you know, because, you know, in, in higher ed, and probably most places, but I think more so in higher ed, you know, your customer base, if you will, changes by 25% — 25% of your customers leave you, you get 25% more customers every year, right? Your seniors graduate, the freshmen come in. So, you know, we're a school of about, let's just call it 4,000. A thousand seniors leaving, a thousand freshmen come in, and they all have the same questions, you know? So you got a thousand students asking the same 50 questions, which is 50,000 questions — they're answer shopping, trying, trying to, trying to find the right answer. You know, with this notion of a generative AI chatbot, I remember the day we settled in on it, but it had been really gnawing at me as to how we could, we got to find a better self-service way for answers. And whether, you know, I don't, honestly don't recall if I got the idea from an article or a webinar, but when it happened was, you know, one of our board members said to me, "Chuck, I got to introduce you to one of our tech guys at alliantDigital. You know, he's really good. He's really, got to talk to him." So, so we, we set up a call, and we got on it, we got on a Teams call. And, you know, Chris said, you know, "We're starting to work on, on, on, on chatbots," kind of in this area that we were both collectively talking about. And he's like, "I could probably demo it for if you send me some data." Like, you don't have to ask me twice to send you data. So we, I don't think we hung up the call I'm like, "Here, take this," and I sent him links to our employee handbook, our policies and procedures, student handbook, and pieces of all of our external data. So it was a very safe way for me to share stuff, because I was sharing stuff that was already out on the internet that we probably wanted people to find, that they couldn't find. Right? So, you know, I sent that over to Chris, and that's where the idea got spawned, is really from that, that first call that we had first. I think, like a week later, I'd have to check my calendar, but let's just say a week or two later, we're, we're seeing a pilot. We're seeing, we're seeing a prototype of this thing.
Chris Stephenson 05:46
I remember the day as well. It's, generative AI has been, it's been in the news for quite a while now. ChatGPT is not brand new anymore. I think it's about to turn two years old, actually. But we've been playing with, with generative AI, and all of the large language models in a bunch of different industries. And when I, when I talked to Chuck, what really jumped at me was, was a couple things that made us really jump to action. First of all, the breadth of information that we had to bring in was so cool. There was multiple websites that had, that updated frequently, that we had to make sure every update was captured in the training model. So it was continuously, a continuous updating language model first of all. But to have, to have students, the, which I think were probably the heaviest users, if, even if you look at ChatGPT statistics, in the spring, after school ends, the usage goes down, and in the fall, it goes back up. So to have our heaviest user base of generative AI really helping shape a direction that we could take, take this chatbot and all of the features and technology that we started to build around it, was really exciting for us. And so yes, when Chuck said stuff, we prioritize getting that, that first version trained up very, very quickly, and making sure we had something good to show.
Rhea Kelly 07:11
I'm just curious how the technology has changed with the advent of generative AI, because chatbots, I feel like, are a really good example of something old becoming new again. Like, what's possible now, that didn't used to be?
Chris Stephenson 07:24
Yeah, it's a great question, and there's been almost three different generations of chatbots at this point. The first was very decision tree based, where you'd ask the question, and if there was an answer to that question, it would it would answer. And if there wasn't, it wasn't. But you had to kind of build every question and answer out in your code. The second was really around natural language processing, and it started to understand the intent of questions, but, but you still had to have answers to those intents. So the, the range of questions that could have an answer brought in, but, but there was still, there still had to be a one-to-one ratio. What's, what's changed now is generative AI allows you to upload content, as Chuck was referring to, such as your website, such as your pretty much any document, and, and generative AI can both understand the context of the documents you upload, as well as the, as the, as the content of the questions that are being asked, and it can match those two things together to give answers. One of my, I'll give two examples, just to kind of show how this has changed with this range that are, I think are pretty cool features as well. The first one was when we, when we first rolled out Ask Tupper, the, we learned that every building at Bryant has a nickname — not the names that's on the website, not the names that's in their, in their handbook. And so when we were asking, "Where's a building?" it wouldn't know the answer. In old chatbot technologies, we would have had to recode every single answer that had the wrong building name, or put, put two, two answers down. It would have been a lot of work. We were able to take one document that was really a cheat sheet of nicknames. And so we took all the real names at the university and all the nicknames, we uploaded that into our, into the GPT we built for them, and at that point, it was able to translate everything afterwards without any issues, and understood both the real name and the nickname. And that's the power of generative AI. You don't have to linear code anymore questions and answers. You simply get the content into one base, and this technology is able to really understand that context and bring the content out. The other one that I really loved was, I'm amazed at how many websites universities have. I never really thought about it until I started visiting them all. And I remember that the athletic director asked a question of, show me everybody that's playing this weekend, and, and the old way to do it was to go to every team website and look at the schedule and write down when they, who was playing and when they were playing. And with, with Ask Tupper we were able to ask that question and have the answer in 10 seconds, because they could cut across all of that content. It was all within one, one base of information. And it was able to answer that, and pretty much do all of those, those clicks that, that used to have to be done manually. So the ease of, the ease of updating information and adding information is probably one of the greatest features that generative AI has really built out. The speed at which the answers come is also quite, quite strong. But the underlying large language models, because they're trained to understand our context and our language so well, it no longer requires an answer to be written for every single question. And that's really where we're seeing the value.
Chuck LoCurto 10:31
You know, literally just a few days ago, I was going through it, you know, we get a chance to see all the questions and answers, and I was scrolling through the questions because I wanted someone in one of the divisions to QA the answers, right? So here's the question, here's the answer it gave. You probably shouldn't rely on Chuck LoCurto, the IT guy, to validate that the advising question was correct. So I had to go through it, so I was paging through, it was a couple hundred questions over a week or two, and I saw a question, and it said, "Where is so-and-so's office?" And it got the answer correct. And I looked at it, and I said, there is nowhere, anywhere in our data, do we say where this person's office is. But this person worked in advising, okay? And I had provided written directions: Where's the advising office? Because one of the questions we get, anybody from higher ed right now is listening, questions you get all the time is, where's this building, where's advising, where's counseling, where's, you know, where's, where's the Writing Center, where's this, where's all that. So I gave, so I usually walk out my office. I walk out to the Roto. It's called Roto. Look, you see that stair? See the hallway there? Go down that hallway, go down the stairs. You get down to the bottom. Make a right, and advising is on the left. So this lady worked in advising, and it knew she worked in advising, but was also smart enough to know that it had directions to advising. So it didn't necessarily, to Chris's thing, we didn't have to have an answer for where's Chuck's office, where's Chris's office, where's Kristen's office. We didn't have to have all that. It just, it just knew. And while that probably seems simple to some, I'm sitting here thinking, that simple, that little thing was so cool because it's starting to understand, right? Based on the questions people ask, the answers and the data that we have, it's starting to put two and two together. It's really pretty cool. It's like, how did it do that? So I took credit for and said, yeah, I programmed it.
Rhea Kelly 12:41
That's super interesting. Do you look at some of those student questions and maybe take away, like, oh, this is information, and we need to disseminate in a better way if, like, a lot of students have the same questions about something in particular?
Chuck LoCurto 12:55
Yes, we absolutely do that. And Chris mentioned earlier what we did with, like, I'll call it nicknames, but we also had to load in some jargon, right? So I used the word the Roto a minute ago. Well, that's the Rotunda in the unit structure. And, like, people just call it the Roto. And the cafeteria is called Salmanson, but everybody calls it Salmo. So if you asked Tupper, where's Salmo, like, it wouldn't know. So we fed in all the translations and jargon and just sort of a cheat sheet to help it sort of put two and two together. So we definitely do that. And then I was a guest lecturer in a marketing class, and the class hadn't started yet. I'm like, okay, guys, you heard we're rolling out Ask Tupper, so it's not released yet, but come on, let me, I'll give you an inside scoop. Ready? Alright, somebody ask me a question. And they said, what's the hall number for the Cumberland Hall? And I was like, what, like, what year are you? He was a sophomore. Like, four years ago, we got rid of the numbers. Why would you need to know the number? Well, there's two reasons: because Facilities people still say, "I need to go to Hall 3 for a broken pipe," and Uber Eats and DoorDash knew the dorm numbers, because that's what was on Google Maps — the names weren't there. So we loaded that cross reference list in there. So we absolutely look at some of the questions that the students are asking, and then, like, a lot of, a lot of the answers are institutional knowledge, like they're not posted anywhere, you know? So getting back to the dorms, so I sent an email out to the community, right? All faculty, staff and students. Ask Tupper's coming, you know, get excited, blah, blah, blah. I go to the gym and I type in my ID, and my name and, and face comes up at the front desk, and the girl looks at it. She goes, "Didn't you just send us an email about, you know, about Ask Tupper?" "Yeah, that's me." I'm like, "Did you ask any questions yet?" She goes, "No." I said, "Well, log on right now, ask it a question." She goes, well, I said, "Well, give me a question. Give me a question I can Ask Tupper." And she goes, "Which dorms have air conditioning?" I'm like, we absolutely do not have that anywhere. I know that stuff is not posted, right? So I went to the head of Residence Life. I said, "Do we have any sort of document anywhere that shows, like this particular residence hall has washers and dryers, study rooms, ping pong, pool tables? She goes, "Oh, yeah, I keep that in a separate sheet." I'm like, "Can I have it? Do you mind if I put that into Ask Tupper so everybody else can get the answers?" So those are sort of the things we were uncovering, making some of that institutional knowledge available to, you know, to everyone — faculty, staff and students. Because right now, we've had it locked down in Active Directory, so only a community member with an ID and password can actually log into Tupper and ask questions. We don't have prospective parents and students yet, until we really get confident in the answers, that he's not going to hallucinate.
Rhea Kelly 16:04
Sounds like a pretty interesting exercise in hunting down obscure sources of data that are around various people's files.
Chuck LoCurto 16:14
But wait, there's more, right? So you look at some of these answers and you're like, what? Where did it get that from? Right? So I got probably dozens of those. You know, one was, we were chasing down this whole thing about the name of the dorms, and I was like, what's the name of such and such hall? And it came back, I forget the name, let's just say it was called Jack Stein. I'm like, what? There's no such thing as that. Is this thing hallucinating? Well, what we, what we fed it, which we since took out, was some, it's called our Digital Commons, our archives. It was the name of a hall when Bryant University was in downtown Providence, on the Brown campus. That was ancient. So Tupper was correct, but his answer was like, 40 years old, right? So we found out, we found out we had to get rid of some old stuff. And what we'll probably do is create a, I haven't asked Chris yet, but we'll probably create, like, an archive Ask Tupper, like, you know, some of the, some of the old stuff. So what, so what we've found is, and I'm sure other universities, for sure corporate America is a little bit tighter on this stuff. There was old PDFs out there about how to do drop/ad. So I asked it, you know, I asked it a question, which was, when is drop/ad over? How do I drop out of class? And here comes, here comes the answer from Tupper. Go to the website. It's got the link, fill out this form. We're like, what? We got rid of that form a couple of years ago. Well, I think it was either Advising or the Registrar didn't know that that PDF was still posted on our website. So we got rid of that. So this was one of those sort of intended byproducts of this, is we knew we would find old content that needed to be taken out of the environment, and we've been finding lots of it. So that's another way which Ask Tupper has been getting smarter, if you will, more accurate.
Chris Stephenson 18:21
The analytics are great on this. I think that's one great scenario where, where this is content that people can find on a website, right, in a document that they're getting misinformation on. By having all the questions and answers centralized into one place, now all of that information is staying current. One thing we're really excited about is, is to open up Ask Tupper to recruits and, and, and to the outside. And in our, and the Admissions team is, is really excited with us, because they feel that search engine optimization, brochure messaging, questions and answers that they put on the website, can all be influenced by the questions that are coming up, in from incoming candidates and incoming visitors. So the ability to not just answer the questions, but then summarize all the questions that were asked, see those answers, and see the count of them, can really influence messaging and investment going forward for universities as well.
Chuck LoCurto 19:18
You know, one of the, we did a few things to get the student body to know about it. Just to be honest, it needs to be more widespread and it needs to be used more. But the way we got it launched was we have something called Welcome Week at Bryant. So two days before all other students move in, the freshmen come in, and they are taken through two days worth of how to be successful at Bryant, you know, where it's, you know, where are all the facilities, and we made sure that there was a section in there about how to use Ask Tupper. And we believe a lot, many of our questions are coming in from those freshmen, and they're really, they're really good questions. I should probably pull up, while we're talking let's see if I can hunt down one of those spreadsheets and pick out one of the questions that I can maybe share later if we get to it.
Rhea Kelly 20:16
So besides uncovering maybe outdated sources of data, were there any other pitfalls that you learned from along the way, or just, or general lessons learned?
Chris Stephenson 20:27
Yeah, there, there were some really interesting challenges with [university students]. First of all, they're very good coders, and I will say one day during the test, they did get Tupper talking like a pirate. We're still not quite sure how they did it. I wish whoever did it would come clean, because I want to hire them now. I want to hire them on the spot. But, but yeah, for about, for about four or five hours, Tupper, Tupper was talking like a pirate to everybody that would, would talk back to it. So learning a little bit about just how, how prompt engineering can, can alter a chat bot, and making those adjustments was definitely one thing. On a more serious front, one of the things that was so important that Brian and I think is just so important in the world, is mental health, and really making sure that, that this chat bot was answering questions neutrally and safely. And we spent a lot of time really thinking that through, and actually used the Bryant team that, that is trained on, on handling issues with students when they, when they do have an issue. So we used the same group that uses their call center line for, for the, for students that are struggling. And we had them really train and give us answers to make sure that when questions came in, when the tone came in of sadness or depression or frustration, that the empathy was, it was in the bot, and the bot was getting them to the right resources right away. We also this year have run into a political, it's a very big political year, and the right and the left have very different opinions. I live up in New England as well, so we're very left there, but we had to spend a lot of time on the neutral response as well regarding questions about anything around politics, anything around where opinions could be strong. We wanted a, we wanted a bot that would answer questions, we, not, not one that took sides. And so building that persona out and really refining the, the responses and the code to recognize different moods, to recognize different tones and questions, and make sure that answers kind of stayed in the, in the Q and A format and not the opinion format, took quite a bit of practice, took quite a bit of training, and really took some experts at the university and our team to make sure that we got that right.
Chuck LoCurto 22:51
Yeah, and that whole sentiment analysis, that came up early on. You know, one of our librarians was part of our initial team, and librarians are typically getting asked lots of questions, so, you know, so some of the not so fun questions, like, you know, I'm feeling a bit depressed. What should I do? Right? So it makes sure that it provides, you know, the number for Counseling or Department of Public Safety. It's doing a very good job of recognizing questions that are, hmm, yeah, better give them a phone number to call, right? But I'm looking at, I promised I would open up the spreadsheet, so row 320: If I have a guest that's not staying overnight, do they need to get a visitor pass? Yes, visitors to any department on campus need to pick up a visitor pass at entry consult, entry control station, and they can park in Lot C. Like, holy cow. Like, there's probably no one that really knows that, except maybe one of the DPS agents that knows, you know, here's, here's what you, here's what you do. So, you know, here's one. What time does Salmo stop serving breakfast? I didn't realize this — 10:30 a.m. they stop serving breakfast. So it's not like they close, but you probably can't get an omelet after 10:30 because they're, they're getting the grill ready for burgers and stuff. So it's all these kinds of, all these kinds of questions. This one has a long answer, but how do I book an appointment with my advisor? Right? It kind of goes through and tells you all the steps.
Chris Stephenson 24:25
That's a good one, Chuck, because one thing we have learned, both in higher ed and all industries, is, for some reason, chat bots are a more trusted advisor at times than managers or human advisors to some of the next generation. And it's, it makes sense to me, right? This is a generation that grew up with phone in their hand. You had a recorded VCR before, don't even know what a VCR is anymore, but knew how to record, knew how to work with technology, sometimes before, right, before anything else. But we're seeing questions about: I'm having trouble with my teacher. What should I do? How do I write a paper? And these questions, as we, as we show them in summary, they're, we're hearing feedback that we're never getting these questions live. And so one thing I think is really cool about this way of communicating is it's giving another outlet for someone that might be too shy, too embarrassed to ask a question into the real world, but it's a way that they can get help. We know during finals, for example, the Essay Center, the center that helps write essays, was probably the heaviest volume it's ever had, because anyone that asked about how do I write an essay would get a link to that resource at the bottom of the answer. We're starting to bring Tupper into the classroom and actually being almost an assistant that can answer questions about a class or about a topic, 24/7, but what we're, you know, what we've heard from that on the professor side is, is that being able to see the questions that students are asking when they are, when they are talking to the assistant allows them to alter what they're teaching in the classroom. And they see a topic is being asked a lot, they know they have to spend more classroom time on it. So the ability to get, I guess, a different feedback channel from students on what they need to learn more, what they want more of, is, is, was an unexpected benefit. We don't just see more questions. We're seeing different questions that are coming in through the Ask Tupper interface.
Chuck LoCurto 26:28
So here's one that helps my side of the world, because we loaded all of our, all of our IT website there, and is the first time I'm seeing this question: What WiFi should my printer use? Here's the answer: Currently, Bryant University's network policy does not support personal wireless printers in the dorms for security reasons. Instead, you could use Bluetooth. If you need a USB cable to physically connect your printer, you can purchase one at Laptop Central in the Bello Center. Of course, that's our help desk, and that's the building that it's in. So it says, or you can just go to Amazon or Staples or Target and just buy one. So that was an awesome answer, because we do not allow printers on the wireless network. We have plenty of printing capability all across campus. But if you want to, if you want to physically hardwire into your laptop and print, that's fine. Because, you know, all the students can see those wireless printers and send all kind of photos to anyone's printer that they want.
Rhea Kelly 27:28
So especially when, with the, you know, starting to incorporate it into classes. What has this, has this sort of led you to need to develop a broader AI policy and, or, what, how are you handling that side of things?
Chuck LoCurto 27:46
Yeah, so, I mean, we established our policy in advance of beginning work on Ask Tupper, because I wanted to make sure the policy was in place first. Now the provost has one that's a bit specific toward, you know, teaching in the classroom and, you know, student honesty with, with having, you know, generative AI write its papers and stuff like that. So we, you know, we put the, we put the policy in place ahead of time, you know, in addition, in anticipation of that. You know, our provost and I work really closely together on this. And I've actually done his pitch and he's done mine. You know, from the academic side, we're changing what we teach, we're changing how we teach. So that's what's going on on the academic side. And from me looking at operational efficiency, you know, how could, how can we, this is an old phrase, but do more with less. But really, that's, that's, that's truly the case, you know, we're certainly not getting any additions to staff, so how can I, how can I, you know, use this tool to help us be more efficient? And quite frankly, this WiFi question is a great one, because that never made it to my team and they didn't have to compose an answer. It actually pulled, it assembled the answer from my policies. You know?
Chris Stephenson 29:05
One policy we see consistently with universities, and Bryant's no exception, is, is to not use AI to take tasks, to not use AI to write papers. That's been a big, that's a worry we hear with almost every meeting we have. And, you know, unfortunately, fortunately or unfortunately, OpenAI, they give the answer. When you ask a question, they give the answer. One thing we've done with Ask Tupper, we gave it a persona of more of a Socratic teaching method, and it will not give the answer to a student. So if you put in a question to a test or a question to a problem, it will not just say, here's the answer, but it will give you an explanation of how to do the answer, and it will always finish with a prompting question or two to say, what do you think about this? Or what do you think, right? What do you think this part of the calculation is? And so we're, you know, we've kind of nicknamed it "Strategy Guru," but, but the goal of it is to help students learn as they're, as they're doing the answers with prompting and teaching. And it'll tell them when they're right, but it will not necessarily just give the answer out. And that's something that was really a direct, direct benefit from, from all of these policies of not using AI to write papers and take tests, necessarily, and a way that we're really, you know, trying to use AI for good in the education sector. We know students all learn different ways, and it's tough for the classroom of a lot of students to teach all those ways completely. This, this Tupper can start to customize the discussion that each student is having so they can learn at their pace, they can learn with their questions, they can learn what their prompts. And so that's something that we really did spend a lot of time on to really adopt that AI policy of no AI for answers.
Rhea Kelly 30:47
Okay, so one last question. I want to hear what's next, or what's on your wish list next for AI? It sounds like some of the things you've mentioned are using Ask Tupper for prospective students, and also the sort of tutoring aspect that's in development. What else?
Chuck LoCurto 31:07
Yeah, so definitely releasing Ask Tupper to a broader group of constituents, potentially have it wide open. But I was just starting to explain to Chris, I think we can use, oh, actually, you've got two things going on. I think we can use Ask Tupper to really do the kind of things with the applicant pool that just could never be done with the amount of staff that we have, right? So I'm gonna leave it at that. So we're gonna do more there. The other thing we're doing is, I forgot about this one, Chris: Pepper, right? How did we forget Pepper?
Chris Stephenson 31:50
I didn't forget Pepper. I was just letting you announce Pepper because I know it's your favorite thing.
Chuck LoCurto 31:54
Yeah, we have a, we have a humanoid robot that, you know, the professors use this in academics, for teaching. And at times we would have it at Open House, you know, walking the floor. It's about four feet tall, arms, legs, eyes, talks, all that kind of stuff. But you had to actually, to Chris's opening conversation around, you had to program it. You know, here's the question, what's the answer. Here's the question, what's the answer. And that's like a nightmare. So we had a brainstorm about two weeks ago. We were having an event on campus, and we were doing tours of our new building, and we were doing tours of the AI Lab, and we keep Tupper, we keep Pepper in the AI Lab. And it just got everybody's attention, because, you know, he's dancing, you know, he's answering questions, and I'm like, we totally need to make this physical robot the user interface to Ask Tupper, because I should be able to ask Pepper the question. He should be able to make a call out to Ask Tupper, return the answer, and have it come out, come out the, voice wise. So we are shooting to have something workable for actual Open House and our board meeting third week of October. We only started, like two weeks ago.
Rhea Kelly 33:17
Wow, that's pretty soon.
Chuck LoCurto 33:19
Yes.
Chris Stephenson 33:20
Yes, it's always quick with Chuck, I will say that. When we move, we move. But the, the idea of having multimodal ways of communicating with, with this new technology is really cool for us. The chat bots are, as you said, they've been around for a while and, but getting to voice command, and, and then even beyond that, getting into hardware where, with robots, and being able to put the software that you use, use for everything, into that hardware, is really, is really going to be a very interesting way for communications to change. This can happen just about anywhere on campus now, and really in any talking piece of hardware, we can put the questions and answers. I think a couple other areas that we're really interested in, in the higher ed space is we're continuing to work with the, with, with, with the digital tutors and, and really helping teach, and learning, what's working for all different areas, is really important to us. The, I'm excited to see the recruits start to use this, because I think as we see, as we see students engage with school and see the questions and get the history of that questions, we think we can actually optimize the entire experience, from first, first learning about a school all the way to deciding to apply and attend there. And so now that all of the, think about all the interactions that have not historically been recorded that are now going to be able to be part of this Q and A on this chat bot and the analytics we can do with that. We're also interested in business development side of the fence. There's, there's, it's very tough for business development teams to reach out to everybody, all alumni as alumni groups get bigger and bigger. We think there's really interactive ways that that a, an Ask Tupper-like chat bot could interact with them as well, answer questions, engage. And so these are some of the areas within the university that we're, we're starting to think about extending into and thinking about how this technology can really, can really continue to accelerate both the efficiency and just the interaction points as well as the analysis of those areas of university as well.
Rhea Kelly 35:32
Thank you for joining us. I'm Rhea Kelly, and this was the Campus Technology Insider podcast. You can find us on the major podcast platforms or visit us online at campustechnology.com/podcast. Let us know what you think of this episode and what you'd like to hear in the future. Until next time.