Use of AI Chatbots

Today, we're diving into a hot topic that's reshaping how we connect with prospective and current students: the use of AI chatbots in higher education.

I’m excited to be joined by two outstanding professionals who have firsthand experience implementing and using chatbots to enhance the student experience.

In this conversation, we’ll walk through the journey of AI chatbot implementation:

• Why these institutions decided to explore AI tools and how they selected their chatbots.
• The process of implementation, from vendor selection to overcoming challenges.
• The current successes and insights their institutions have gained.
• And most importantly, how you can apply their experiences to your own institution.

Host:

Dr. Alex Fronduto
Faculty Lead, M.Ed in Higher Education Administration & Associate Teaching Professor 
Northeastern University

Guests:

• Devin Purgason, Associate Vice President of Student Experience, Marketing, and Outreach at Forsyth Technical Community College.
• Lenell Hahn, Director of Admissions at Southeast Missouri State University.

Transcript

Copy Transcript Copied

0:09

Welcome to the admitted podcast hosted by Acro. This is Dr. Alex Verdudo, host of the podcast. In this podcast episode, we are going to be diving into a hot topic that's reshaping how we connect with perspective and current students, the use of AI chatbots in higher education. I'm excited to be joined by two outstanding higher education professionals who have firsthand experience in implementing and using chatbots to enhance the student experience. The two guests are Devin Perguson, who is the associate vice president of Student Experience Marketing and Outreach at Forsyth Technical Community College, and Linnell Han, who is the director of admissions at Southeast Missouri State University. In this conversation, we're going to walk through the journey of an AI chatbot implementation, including why these institutions decided to explore AI tools and how they selected their chatbots, the process of implementation. The current successes and challenges. Insights on what they've gained through the process of implementing it and how they're going to continue to explore and Continue on with the process, as well as most importantly, how can you apply their experiences to your institution? Let's start by getting to know our guests and welcome to the Admitted podcast. Hello and welcome to the Admitted podcast hosted by Acro. My name is Doctor Alex Rondudo, and I'm one of the hosts of the podcast. I'm delighted to be joined by two extraordinary guests who are gonna chat about the use of AI chatbots. I'm joined by Linnell Han, who's the director of admissions for Southeast Missouri State University. As well as Devin Ferguson, who's the associate vice president for student experience, marketing and outreach at Forsyth Technical Community College. Welcome, Linnell and Devon.-- How-- are Fantastic. Thank you. Thank you so much for being, of course, thank you so much for being here. Well, before we get started, obviously I want our listeners to know a little bit about each of you and maybe some context around the institution you work for if that's gonna be helpful as we gather our conversation around kind of AI chat pots and the uses of it. Um, so Linnell,-- I'll have you start and then after that we can go right into Devon's-- intro. Yes, I have been working in the world of admissions for 21 years and um. Gosh, our institution serves about 9000 students, and that includes, um, domestic students, international, undergraduate, graduate, completely online degree seekers. We've got regional campuses and so, and we really are working with a multitude of student populations to help them meet their career goals and match them to the right degree for that. Awesome and uh I have been working in higher ed marketing um for almost a decade. I've been at Forsyth Tech for 4.5 years, and Forsyth Tech is a community college here in North Carolina. We have a 2 county service area, so we serve Forsyth County and Stokes County. It's a very regional school. Um, they're actually 58 community colleges in North Carolina and so we're, we're one of the hashtag great58, um, and, uh, we, we, we love our community college system a little bit about our, our student population. We've got about 9000 like curriculum students, so those are students that are seeking an associate's degree. And then whenever we look at our continuing ed program, so you know, programs that are more uh diploma certificates, uh, you know, 12 week short term courses, uh, we have about 22,000 students that we serve, uh, every year. So, so, uh, we, we are very, very, uh, we, we love our community, uh, we, we are here to serve the unique needs and, and respond to those of our, our unique community, so, uh, regional college here. Well, thank you both again for intros being here and do everything you've been doing in higher ed for so many years. So thank you, first of all and foremost. Well, our conversation today is really gonna revolve around AI chatbots, so this will obviously is in both of your wheelhouses in the admissions, recruitment marketing worlds. And really what I hope our conversation is gonna entail is a little bit about the process of even thinking about doing this all the way to where you might be today in terms of implementation, maybe even some success stories. Ideally, the goal is right, others may be thinking about this. Some institutions may even already be maybe in contracts or going to be in contracts with maybe companies or maybe trying to build this something in-house. So I think a lot of people are going to be interested in this topic and so again thanking both of you for coming on and being able to just share your thoughts. I thought I would start first for each of you to kind of give a little bit of the context of where you started and then we might kind of go back and forth in terms of kind of each of your stories because I'm sure there'll be similarities, but I'm sure there are also differences. But first and foremost, let's start with where did the thought process even come from in terms of when you wanted to start implementing it. Well, I'll start with you. So for us, um, our ability to utilize automation really started in 2020, um, and then AI tools really started hitting generative AI, right, hit the marketplace in 2023. And we started having conversations with our ed tech partner about tools they were gonna be adopting and rolling out in April of 23 and we immediately said, yes, we want to be an early adopter um and be right there along with you as you roll out these tools and You know, we wanna make sure that those that were serving, that they have the same types of experiences for other things that they're doing online and it's just, it's becoming an expectation, right? That you've got someone that can assist you 24/7, 365 days a year. Um, but then it's making sure that you adopt the right tools that are gonna be able to complete the items or provide the assistance that's needed. And so those were really key factors, um, in making that decision to, to partner really early. And what were you using prior to, so you kind of alluded to the fact that you might have been using some automation or AI before, you know, Gen AI. What were, what were you already doing? Yeah, so, um, the automation that we had really wasn't conversational automation, but we did have that traditional kind of rule-based chatbot to where someone could ask a question and get an answer, but if they didn't ask it in the right way, let them know that, oh, you know, you still have to actually call the office when they're open Monday through Friday 8 to 5. OK, OK, but no, but you still started somewhere. Some schools never even got to that point. So I mean they're still, you know, working in that direction where it seems like you are a pioneer, you're asking to do early adoption of things, which I think is so crucial because I do think some institutions are pretty hesitant about those things. They're like, oh no, that might hurt the student experience or what if it doesn't say the right thing or gives the wrong information, right? There's so many things that people might be. Thinking about and so I, I applaud you for just saying no, we should do it, let's try it. um, so how about your case, Devin kind of where did it all come from? Yeah, very, very similar to Linnell. um, we, we also, um, had, uh, you know, a couple of what I'll call dumb chatbots, um, you know, we, uh, we had specific ones that were, were. Through a financial aid provider, so it was nothing that was generic or general where a student could go and ask, you know, any question. It was mostly just, um, financial aid based. We had a library chatbot, you know, that was, that was heavily used, yeah, yeah. So we, we had a library chatbot on our website, but we didn't have anything that was nearly as um As impactful or had the breadth that our current AI chatbot has, um, you know, very, very similar to, to Linnell we um our ed tech partner we have the same tech partner, um, and they came to us and said, hey, we're, we're launching this, we're, we're looking for beta testers. Is that something that you would like to do? Our president at Forsyth Tech was incredibly AI forward, um, you know, really at the inception of OpenAI and Chat GPT coming to market in 2023, she decided that we were going to have an AI task force. Uh, so this was something that She was passionate about and wanted to make sure that we were on the cutting edge and that we were not falling behind with this. So this fit in perfectly with that strategic um alignment with our president with with her vision, uh, so this was it was it was fantastic uh fantastic opportunity that we that we had, you know, I didn't, I didn't have the, the headache of going to market and You know, searching for, for who whoever or comparing vendors or anything like that, our CRM uh student engagement platform, they, they wanted to really invest in this, um, this AI focus, this generative AI focused um vein and that's. That's how we uh got our AI chatbot. I love that you talked about the buying piece, which was actually right what I was thinking about, right? You're both like, yeah, we definitely want to do this. I assume you can't just make every decision. We all kind of report somewhere in our, in our hierarchy of like the final decision maker. So Linnell, talk to me about the decision making for yours, like, did you, who did you get buy in from? Was it easy? Was it more difficult, um, within the specific, you know, this new pieces? Yeah. Um, for sure our, our chief information officer, our chief marketing officer, our chief enrollment officer, um, we're all a part of that conversation of course for our information chief information officer it was making sure like the security of our information and data that the AI was gonna be sharing through these conversations that that was in place that was. Key to them, um, from our chief marketing officer, of course they wanted to brand the bot and make sure that it looked very visually pleasing on our website and then our chief enrollment officer is, is, is the content they're gonna give back to students gonna be accurate? Is it gonna be correct? um so everybody had an area that they were really concerned about through the vetting process. I completely understand that and those are all things that I was thinking about, right, in terms of especially accuracy and branding. And so one thing I thought that you were just saying that was interesting is the data security piece which I'm sure what everyone is thinking about. And so in both of your cases, I know that you have the same vendor in some ways, but. Where did the information come from? Did you have to give them a ton of data, obviously, because the input is so important? Did they take it from somewhere? Is it a combination of both? Like, how did you be able to set it up for anything, right? Cause ideally, I assume they can ask anything to it. So where did you kind of give all the inputs? I mean, it's definitely a combination, right? So because it's tools within our CRM it already has that rich data on the student's records, so it has the ability to do things like check on their application status or sign them up for an appointment or a campus visit or tell them what documents they're missing with their application, which is Amazing. Yeah, and then, um, you know, Element 451 has actually already like loaded into the knowledge base too like fast food.gov. So if they have questions about the financial aid process, it can give those responses and then we have our own knowledge base too that we curate as an institution that specific. to us whether it's learning our web pages, uploading documents, linking our YouTube channel so it can return video responses, um, and so you have this sense of security that it's not pulling from the open web, right? It's, it's peering from, it's pulling from a knowledge base that we've curated and we maintain and skills that we've authorized it to have. OK, and that's exactly what I was trying to get at. It's like, where did it all come from? Um, similar Devin, different experience. Yes, very, very similar in the fact that, you know, all of this data was already within our CRM, which is very vetted for security, data privacy, SOC2, all of those acronyms that the tech people that are listening might understand and that I just know enough to be dangerous. With hey yeah, you did a lot of good rattling right there, so I believed you. That's right, yeah, yeah, see, I, it's all mad it's jazz hands, um, and so, uh, you know, so that part was, was already, you know, solidified. Foundational. Where we started to on board and create the chatbot to personalize it to forsyth Tech was through that knowledge-based curation and I think something that we saw was that there was a human resource need for that. Um, When We started to compile the knowledge base, we realized that, you know, knowledge. Some it it it depreciates. And it becomes incorrect or it becomes obsolete, or it becomes wrong, or, you know, if you, if you have your chatbot with The start dates of your semester for the last 10 years, there's gonna be some confusion there where it might pick up a date from 2017 and so making sure that there was someone who had their finger on the pulse of the knowledge, making sure that it was correct, making sure that when, whenever it said something incorrectly that they could then go in and Add a a source to the knowledge base or create a custom instruction to say, oh, don't say that again, or or something like that was was really really important to us, so. Um, we, we actually hired a position to manage our knowledge base, uh, and she not only manages the knowledge base for the chatbot, she also manages the knowledge base all of our knowledge bases for, you know, our student care representatives who are very much like a customer service, um, representative. You cannot absorb all the information about every single area of the college and every single program you have to have a good knowledge base that people can search and can look for, uh, and so that we, we had to add that human resource component which I think is really interesting because I feel like a lot of people feel that AI is gonna take all these jobs away, uh, where we we saw where it it actually has helped us. Not only um have we had to hire people to maintain it, but it's also allowed the people that currently work to have more impactful work with students. So instead of our student care representatives answering the question of when does, when does the spring semester start. Uh, a million times a day, they're able to actually have a conversation with a student that might have an issue with their financial aid, uh, and that's gonna inhibit them from coming the next semester. So, uh, we, we've just seen it where it's, it's really helped us, um. be very intentional with our students. I love that you talked about information and you talked about kind of the capacity and how work is changing. And so I kind of have two sub questions. So the first part that I really wanted to ask about is the knowledge base. So you both touched upon the idea that you need the knowledge base. I don't know if every institution has a knowledge base other than their website, so I'm curious, is that what you did? Did you just like upload a bunch of handbooks and the website into it, or did you physically try to capture more information that was not present on these items? And Linnell, feel free to hop in to kind of talk about that piece. I mean, oh yeah, for us, it took us a matter of hours to get like our version one of our knowledge base together because we really did focus on it learning our web pages. But then, as Devin was saying earlier, you can't just set it and forget it. You do have to have a team that's checking on those conversations and looking for those knowledge gaps and then it really was a continual. Um, management of the knowledge base that as we saw gaps, we looked for, well, is there custom questions and answers we need to upload or is there a document we should upload or is it a change we need to make on our website? And so that's been really good for um just making sure content is. is where it needs to be on our website as well as our, our AI assistant knowledge base. Yeah, and You know, Alex, you asked if we uploaded our website, did we upload handbooks, did we create things out of, out of, you know, do we have to sit down and create things? The answer to all of that is yes, we did, we did all of that. Uh, we were very lucky in the fact that when the AI chatbot was implemented. It was on a brand new website. We had, we had my marketing team and I, we had gone through and redesigned our website which was an 18 month process. It was brand new content on every single page that was up to date. So very, very lucky in that fact.-- I was-- gonna say that lines up very well, yeah, because if you, if you do have a website that is a little bit of a beast. Uh, or, or jungle of, of information and outdated information and thousands and thousands of pages. It's probably not advisable just to copy the home page and pop it in there and say, you know, read the entire website uh you have to be a little bit more strategic with that. So we, we were lucky in the fact that we, we pretty much could trust that the information from the, from the website that we uploaded uh would, would be accurate. Then we also did add all the handbooks, all of these written informa these written documents um they've continued to. Uh, roll out all of the different types of documents that you can upload, so you know PDFs or uh Word docs or even YouTube videos so we're able to upload YouTube videos, it will transcribe that video and then it's able to even share that video uh with students if if it's pertinent to the question that they're asking, uh, so. We, we did that but then also one of the pieces of our knowledge management specialist, her, her job, uh, was to go to all of these different areas, you know, financial aid, admissions and records, student success center, all of these different, you know, um, academics. And sit with them, ask them what are your frequently asked questions what are, you know, ask them all of these, these things to build that knowledge base herself and so what she does now is, you know, once a semester she goes, she sets up these meetings with all of these teams, goes through each of their individual knowledge bases with them. And you know, what information do we need to add, what information is obsolete now and then we need to edit or or things like that. So yes to all, yes to all. I love that. I mean that I guess one of the questions I think a listener may have then is how long did the implementation take again, as you both said. In some ways it was quick and in other ways it was a lot of pre-work with the ideal situation being that it would then be less work over time, right? Because if you input it all, you would answer less questions of the common ones at least. So I'm just curious, like, in a general sense, how long did it take you to do all of this information gathering and like actually launch, you know, version one, let's say. So for us, it was very quick in the fact that We, we were really scaling with the AI chatbot. So for the first several months we did not have it where it was talking to students outside of our working hours because we, we wanted to make sure we wanted to test, we wanted to see what, what questions are coming in, do we have information for that? And so we started with very kind of sparse. Information and have continuously built upon that so it it was uh it was a, you know, as far as the implementation of it, it's one string of code, one string of code that we put on our website, you know, we use WordPress as our CMS, um, and it pops it on every single page. And It was a gradual increase of, of how we would build that knowledge because I heard someone say it like this, like the first day that you implement your, your AI chatbot, like that's the most unintelligent that it's gonna be. It's of course it's just going to continually get smarter and smarter and smarter and so, um, a little bit of that. So for us, uh, what we call like our all knowing and general chatbot, I mean it really did just take less than a week, um, for just days. Now, the cool thing about our ed tech partner is they've now rolled out assistance, AI assistance, so you can have these teams that are specialized with skill sets and so those took us much longer, um. And you may be asking like why do you need like a team of AI assistants. I was thinking the same question,-- so please tell-- us. So you know I mentioned earlier, you know, we serve students looking for graduate degree programs. We serve international students and we serve like that traditional. Like on campus residential student, and for each of them, like they may ask the same question, but the answer needs to be different. And so with these AI assistant teams, if it's a web page that's dedicated to international students, They can only have access to parts of our knowledge base that is specific to international students. So you don't get information that gets crossed and incorrect. So like our application fee is different for an international student versus domestic, and so that ensures that when they say what's your application fee. It is a correct answer, um, and then same thing, you know, graduate information is different or maybe information is different for like an online degree seeker like their onboarding process or orientation, and so it's really great to have these specialized teams that can be on those web pages that are specific to them. Those took a lot longer because we actually gave them personalities that we gave them advert avatars that look very human um and then of course designated what areas of the knowledge base they're gonna have access to what skills we're giving them, um, so that has taken us a long time and we're doing like little subteams at a time, um, but that's been a fantastic. Um, way that they've really enhanced our generative AI capabilities to assist students. Yes. So when you're saying these AI teams and assistants, for example, you gave like the international student example, do those students need to be on that page so that assistant is specific to the international student? Let's say they're not in your CRM or like how do those, how does that piece work, the functionality like you were saying? Yeah, so they would have to be. On a page where that team is, and now they can hand off to each other.-- The AI assistants-- can. Oh God, it's like a team just like working on I'm going to get you this other agent to come on in. Yeah, like that warm handoff from admissions to the housing department or something, right? Yes, yeah. And then there's also, I mean, we still have our all knowing, like our general chatbot is still in existence, so they can always step in at any time. And we haven't even talked about the fact that our AI assistants are more than just web chat. They can also respond to emails and text messages. Exactly. Oh, yeah, I mean that's important to talk about. So it's coming, so you are using AI essentially writing tools as well. Let's let's say an email is coming in and someone's responding that isn't a human being or where is that coming from? It's just like if they would be having a conversation on the web, the conversation via email. So let's say we send a text message to a student from our CRM platform and the student responds back to that text. Our AI assistant will pick up the conversation first, um, and assist them. Perfect, that's what I assumed. Um, I know a lot of outside of enrollment management, but a lot of sales, for example, are utilizing those types of tools to, you know, continuously engage people whenever they want to be engaged, right? At the same time, you may not be physically present there and you've already mentioned it way back in the beginning of that, you know, the instant gratification or the instant answer, you know, trying to meet that need as best as possible, um. So, that was so interesting for both of you to talk about the implementation that it actually didn't take you that long. Obviously it continues to build in time you adding more things, but for both of you it seems at least your version one didn't take much time at all. Um, So my next piece was. In terms of the ongoing piece, so some of you said like you started off with just the 9 to 5 or whatnot. I assume now it's 24/7, and how many conversations would you say it's having in a day? You know, that's such a good question. I, I actually had some data pulled up, uh, but it was, it was gonna be historic data from the entire, uh, The entire um Time that we've had our chatbot on, but let me, I, I, I, well, now, how about you talking and I, I'll pull up some,-- some-- recent. I love, I love, I love data. Yeah. I, I have my dashboard open too here, um, with an element and You asked about a day, so let's see. We've had in one day, 222 conversations. That's pretty impressive. I was looking at December 1st, which was a Sunday, which is a big deadline for us, and our, our AI assistant helped 63 families on a Sunday. But hey, that's still 63 people that, you know, we're not waiting, ideally, hopefully they were all solved and weren't waiting until Monday, you know, post deadline. Um,-- so-- I tell you how many of them actually had to be handed off to a human user. I was gonna assume that you can pull that. Only one requested to be to to talk to a human assistant. So look at that, right? Like that's that's pretty awesome. Thank you both for sharing that. I think that will help listeners really put that in context and, you know, this goes along with the second piece, you know, going way back to earlier in the conversation, Devin, when you were talking about like you added a staff member, right? Like people are thinking like, oh, we're not answering the 62 other questions right there on that Sunday, right? Like, we don't need as much admission counselors or financial aid counselors, you know, insert position here. But There's still work that has to be done on the back end that I don't think people are always considering, um, and so it's great to hear that. For many people, and I've said this in a lot of different avenues when we're talking about Gen AI in general, right, it's, it's not necessarily always the positions that are like being eliminated, right? It's, it's the change of work and what is that person doing? What, how is their capacity changing from what they might have been spending a lot of time on, where they could actually, and I love how you said this a little bit, Devin was like, going more in depth or getting more, you know, to the specific conversations that need to happen. I think about Especially in the enrollment and financial aid and thinking about caseloads or territories, right? Like, could you be having more in-depth conversations or more in depth, you know, issue solving if all of these general questions were now out of the way and not clogging your voicemail or inbox or anything. Um, so I just kinda want to kind of bring that back together. Uh, I don't know if either of you had any thoughts around that piece, like, I don't know Linnell, for example, like, do you, did you hire a position? Did it change any types of positions or work? So we did not hire a position dedicated to this, but we did take more of a team approach, um, and we're pulling in several individuals within the team. We're a Microsoft school, so we use Microsoft Teams chat and we created like a watch group for. And so, so there's a group of us that um keep an eye on those conversations and they'll drop into that chat when the, the answer was incorrect or we need to review our knowledge base and see how we can tweak it for it to be a better response in the future. For for us it, you know, we added our one position that was the knowledge management piece. But what it did adversely, When we were building our student care team and our, our contact center. We took the volume of all of our calls and all of the things that were coming in and we, we put it in one of these call center calculators and it said you need 16 full-time people to handle all of this volume and so obviously we did not have 16 full-time positions to just hire and and disappear out of nowhere. So yeah, so we, so we were scaling up toward that. And what it allowed us to do was after it took, you know, uh, last spring, our call volume went down 30, 35%. We, we, we're now at 10 people, very, very well staffed. So it, it didn't. It didn't eliminate positions in the fact that, you know, people weren't having to work. It, it, it allowed us to um scale back because the need wasn't as great. We weren't getting as much of the call volume. But I think one thing that's interesting, it, you know, that doesn't mean that the team is just sitting there twiddling their thumbs. We saw that the call length. The average call time. It was much longer than what it was in the past because the student care representatives are able to actually have a conversation with the student, see what their needs are, and, and help them uh at a much, much more in a much more robust way than they, they were previously um able to. I think that's a great point to think about the average call length and again, looking from the student experience, the student help aspect is that you can just spend more time, you can deal with, you know, the specific situations that you may not always be able to feel like you have time for, um. And so kind of now today and reflecting back, so you both have started to pull out some data in terms of you just said, for example, Devin about call volumes decreasing in percentages of calls. Um, now, what have you been seeing, you know, now that you've introduced it, kind of what are some of those successes you have seen from it? I mean, you know, just the satisfaction of a student feeling like they can get help whenever they need it has been amazing. And because of the fact it's, it's on top of our CRM and that it can check on someone's application, let them know missing documents, those are things that we would have traditionally received in an email or over the phone and so it has shifted for us that volume and We don't want anyone to be bound to their desk all day, just answering emails. We want our team to be out in high schools and at community colleges and meeting one on one with campus visitors, and it really frees up time for us to Be proactive with our outreach and connection building versus just constantly being reactive to questions coming our way because the AI assistants are, are taking that burden off of our plates in a significant way. I love that, again, like you said, what else can you be doing and how can you know, the positions change slightly or maybe just reemphasize in different areas, um, in. Depending on when you implemented and how many cycles you've had and. Do you feel like you have any data points to show that it's has specific increases in any area other than just like, you know, it's reduced calls, but has it, could you say in any way that it's helped with like your enrollment or helped in terms of, you know, retention depending on how you're utilizing it? You know, I think it's really hard to say one thing is what makes the difference, right? With, um, meeting, uh, you know, goals at an institution. I really think it's a combination of strategies and tactics, but this is something that our goal is to increase satisfaction for the students we serve, um, and if they can get their answer quickly, rapid rapidly like on their time and not on our time. Um, it's gonna make them happier through their college search process. Exactly, yes, very, you know, we have seen enrollment increases, uh, we have seen retention and persistence increasing over the last 2 or 3 years, which has been fantastic, but once again, I, is it because of the AI chatbot or is it because Our student care representatives have had, have had more time to really engage with students or our advisers have been able to really dig in and help students and make sure that they're successful, um, because of some of the, so there's, there's a myriad of things, but I, I definitely would say I, I don't want to go back to the time where I did not have this tool. OK, see, that's, that's good. And I, and from like anecdotally, let's say from your staffs, let's say, do have they expressed that they feel like this some of this burden, like you're both saying has been taking off their shirts? Are they also feeling like they're getting less calls, less emails, etc. and allowing them to feel like they have that more capacity that you're saying? Definitely making the day more manageable to, yes. So, you know, we've had a lot of conversations of making sure you're blocking time on your calendar to do those activities, they're gonna be the most impactful. And that 3 to 5 when someone's out of school and they're available, if you're using that just to respond to email and you're not using that to make phone calls, right, then it's really, it's been a game changer for in that regard. Yeah, same exact situation for us. I think our teams have really enjoyed the extra time that they're able to spend with students because of the AI chatbots. I, I think even our staff, they use AI chatbot to ask questions, you know, it sometimes it's faster, it's faster than actually searching on our website sometimes. So I, you know, we, a lot of people utilize it. It's, it's really fantastic, so yes. And then the flip side, you were saying, Linnell, excuse me, that the satisfaction piece so. I know most AI tools, chatbots may have like a thumbs up, thumbs down, or do you feel like your question was answered. So what kind of feedback or data are you getting on that kind of satisfaction? Obviously, the data point you already gave about that Sunday, only one person needing additional service kind of tells me a little bit, but like you said, that doesn't also answer the question of like, the satisfaction around that. So I'm curious, what are you looking? Yeah. So within um Element 451's AI assistance, it puts this like little mood on the conversation. So you can see if the student we're helping is happy, if they're upset, or if they're just kind of neutral. So you do that and you can tell that from the tone too of the words that they're using when they're conversing. And then what's so great about this generative AI assistant is that it I mean, it has feelings like it uses words that is a very complimentary, very nice, you know, I worry sometimes when our human team members, you know, our mood can have an impact on that conversation with the family, but our AI. My assistant is always happy. They are always thankful. They're always asking what more can I do to assist you. Um, they, they're gonna spend as much time as they need to to make sure that they are happy, um, and everything is taken care of. And it also will, it'll use their preferred name in the system, which is really great. It it will, it will use their name in the response back to them. It will summarize what they've said. And it, I mean it really will, it will uplift them, it will encourage them, um, it just makes you feel good, like just reading those conversations and checking in on that, it's, it's awesome. Yeah, I mean, there have been several conversations that I've looked at where the chatbot has been like, yeah, yeah, Claire, you can do this, you got this. We believe in you. It's, it's just, it's been, it's been really wonderful. Ah, that's so great to hear, and I hope that, you know, more people think about, you know, the, the benefits that it will have if, you know, they decide to adopt something similar, you know, as we close out. Obviously there are people here that are listening that may want to think about it, maybe they don't have your CRM, etc. like if you had to give them one recommendation or one way to advocate for it, right, that's usually the number one issue is advocating budget, you know, buying. Is there anything that you would share to those people? The one thing that I would share to those people is, since July 1st, 2023, our AI chatbot has saved 170,496 minutes. I love the data. I love data, so thank you. I mean, if you think about that from a human resource perspective alone, not even, not even considering the student satisfaction and being able to get answers whenever they want it or whenever they need it, just the human resource capital, it pays for itself over and over and over again. What now? Well. It is a part of everything that everyone's using today. Whether they're on social media, you have your AI assistants in there, if you're on your web browser, you have your AI assistant. It's everywhere and so, I mean, why Wouldn't someone who's searching for colleges want to have that type of support and tool as well. So it's where the future's going. I feel like if you're not willing to at least consider it, um, then you're gonna be left in the dust. Well, I appreciate that and thank you both for sharing all of this great information, becoming on the podcast. I so appreciate it, and I thank you both for being here. Thank you so much. Yeah,-- thank-- you.