In this episode of Generation AI, we explore the evolving world of custom chatbots and their potential impact on AI usage in higher education. We discuss the personalized and commercial applications of these bots, drawing parallels to the early days of the App Store. We also highlight the recent launch of OpenAI's GPT store, emphasizing the customization and personalization capabilities of GPT bots. Tune in to gain insights into the development, usage, and commercialization of custom chatbots in the industry.
Ardis and Dr. JC Bonilla, take a deep dive into the world of custom chatbots and their impact on the industry. Drawing on their own experiences, they share their insights into building and using custom chatbots, particularly within the context of higher education.
One of the key highlights of this episode is the introduction of the GPT store, a platform that offers custom bots, developed by OpenAI. Our hosts explore the potential for monetization, discoverability, and user interactions within this innovative marketplace, shedding light on the exciting opportunities it presents for both developers and users.
With a focus on personalization and commercial applications, Ardis and Dr. JC draw parallels between the emergence of custom bots and the transformative impact of the App Store in the early 2000s. Their predictions for the future of AI usage are thought-provoking, as they anticipate how custom bots will shape the landscape of artificial intelligence.
Ardis Kadiu [00:00:06]:
Hello, and welcome to AI, the podcast where we demystify artificial intelligence in the world of higher education. I'm your host, Artis Gadu, and joined with me today is doctor JC Bonilla. Hello, JC. How are you doing today?
Dr. JC Bonilla [00:00:21]:
Hola. Hello, everybody. Hi, Ardis. Hi. Are amazing. I think a few 100 listeners on this podcast. We're growing ardis.
Ardis Kadiu [00:00:31]:
So today, we're gonna talk about custom bots, which
Dr. JC Bonilla [00:00:35]:
Bots are taking over.
Ardis Kadiu [00:00:36]:
They're taking over. So in kind of a world where apps revolutionized how we use our phones, I believe bots are going to reshape how AI is going to be used at least for this next year or 2.
Dr. JC Bonilla [00:00:48]:
When was the App Store launched? Early 2000?
Ardis Kadiu [00:00:51]:
Yeah.
Dr. JC Bonilla [00:00:52]:
Early 2000s, and what we are saying is that, relevant, transformational, And impactful the launch of the App Store was, and it continues to be. What we are about to foresee in In front of our eyes is a deployment of a marketplace for AI. And the premise of today's angle is that Generation AI, we get to see it first, and it will be in a look and feeling of bots, b o t, bots. Right? Fascinating topic. I cannot wait to see it, and let's see how this episode and conversation unfolds. So how do we start the conversation about chatbots, marketplace, an AI next step.
Ardis Kadiu [00:01:37]:
So you use your phone very differently than I do.
Dr. JC Bonilla [00:01:40]:
A lot of selfies. A lot of selfies.
Ardis Kadiu [00:01:43]:
Exactly. But the apps that you put in there are very personal to you because that's how they organize your world. When we think about it in the world of AI. We have the general purpose tools or AI and bots. So CHAT GPT is a bot. Bard is a bot. So a lot of these foundational models, they have built interfaces on top of the Toast chatbots. However, it takes a lot of time to personalize that chatbot to give you answers that you want.
Ardis Kadiu [00:02:09]:
We've talked about it in prior episodes how JC has specific prompts to personalize the output and to personalize what these chatbots are providing. Now what we see in the AI world, the ability to personalize these bots into creating new personas and new use cases
Dr. JC Bonilla [00:02:28]:
Mhmm.
Ardis Kadiu [00:02:29]:
Is gonna be the equivalent of making technology more personal, very similar to how apps made phones personal. Now AI and bots are going to make our interactions with AI a lot more personal going forward.
Dr. JC Bonilla [00:02:41]:
So here's an example. I've spent time developing JC Ethos, which is basically the name of my bot, the one I I put together. There in a way, when I've prompted, I figure out that this is what I like, right, which is a response in the tone of an academic that often runs into a number, but likes to draw history, narratives, sometimes funny, very, very concise, and leads with the opportunity of establish a conversation, right? So all that, it's a series of prompts and it's my tone. Right? If you were to add an accent and make it the person bold, all those type of things, now you go to the next level. But at the very first of all, there's the tonality and the response type. I've packaged that into a bot. And for whatever reason, Artis. You like what I do and how my chat DPT outputs emails? So here is use my bot, and potentially, you can send me, I don't know, $3 every time you put something together.
Dr. JC Bonilla [00:03:41]:
I don't know how commercially we're gonna do this, but that's basically the application here. I've done that prework, and now, eventually, It can be going somewhere, and you can benefit from it. Is that right? Is that how you see this personalization in the phone and whatnot taking place in AI?
Ardis Kadiu [00:03:55]:
That is exactly right. So we are packaging workflows. We're packaging personalities and personas together, and that packaging is going to be a bot. The underlying technology and the underlying models that are being used are going to be commoditized. However, the personality, the use case, the workflow, the additional context that we provide as part of that. And the wrapping up of those pieces altogether constitute what we're calling a bot today. These bots have personalities. They have behavior.
Ardis Kadiu [00:04:30]:
They have tone of voice, and they have different output styles. And to your point, they are the JC bot. You can wrap that up, and now you can hand it off to somebody else, combining all those pieces very similar to, you know, an app in the App Store. We are just wrapping all these pieces together into bots so we can provide them as apps. The one drawback is that the app stores of AI are not available yet. And today, we're talking about OpenAI announcing the availability of their app store. And prior to that, they opened up Pro Chat GPT with their notion of what they call GPTs, which are custom bots, wrapping up those prompts that you mentioned before and making them shareable. So have you used any of that functionality yet, JC?
Dr. JC Bonilla [00:05:25]:
As a user, I've stayed in the domain of creating my own bot and stays within my instance. Right? But taking the next steps to prepackaging or standardizing. I haven't gotten any of that. Wait. I have a bot. You know, let's go to university land. Right? So, you know, the golden eagle, the mascot Of John Brown University, and there's a bot. And they've been investing years in creating that bot that interacts with students.
Dr. JC Bonilla [00:05:52]:
I'm an applicant. I wanna know an event. That bot. So, Arnie, are you telling me that that bot now goes into the marketplace? Is this what we're talking about, or is this different? How is my bot investment of gazillion dollars and, you know, n number of time now entering or not entering this custom bot marketplace.
Ardis Kadiu [00:06:10]:
That's a very good example. One of the things that we do here at Element is provide different bots specifically working on knowledge bases for institutions. And Boldbot is our general purpose bot, which can be customized and has different personalities for different institutions. And what we do is we actually use the content and personalize it to the school itself. So the school loads the content in, or they point it to different data sources that they have, and we're connecting it directly to the data sources in the CRM so it knows things like who the person is, and it has context. It has history. It has real time information so you don't have to go back and train it over and over or load that historical content over and over. What we're talking about in with chat GPT and their GPT store, it can be very similar to what you're mentioning, but it lives in the chat GPT e World.
Ardis Kadiu [00:07:08]:
It's branded as Chad GPT and the GPT bots, and the app store is exactly the same. So you're in the ecosystem of OpenAI. So think about it. Now if you want an app that's only on iOS Yep. But you have Android, Well, you can't have that.
Dr. JC Bonilla [00:07:26]:
Exactly.
Ardis Kadiu [00:07:26]:
So how does that work? If you are in OpenAI world, you have to use Chat GPT in order to get this bot that JC created. Are you going to create it in multiple different platforms so everybody can have access? Probably not.
Dr. JC Bonilla [00:07:42]:
Artist, I don't have the time to do that, my friend. But the reason why this requires you in that whole episode is that in terms of market share, we haven't seen the stats, but we all anticipate that OpenAI's dominating market share. And as such, it's going to probably start accelerating how other organizations start thinking about marketplaces. Right? Do we have any stats, by the way, on how these custom bots are being used and things of that sort? How can we quantify that?
Ardis Kadiu [00:08:07]:
Yeah, which one of the interesting parts around custom bots is that the platform called Character AI, They provide the ability for you to create these different bots and personas. If you haven't played with character AI, I would strongly recommend you do it. So you can go in there, and there is an Elon Musk bot. There's all kinds of different bots that you can kind of play around with. And the tone and voice and how they bond. They're gonna be very similar to those different characters. You can build your own as well. There's folks who are building virtual girlfriends and virtual partners as well on this platform.
Ardis Kadiu [00:08:44]:
You mentioned stats. I believe the average time that folks spend chatting with bots is upwards of over an hour or 2 a day, which some of the people who use it the most, like, they're spending a tremendous amount of time interacting with these bots, and it can be anything from communicating with them, just telling them personal things, to having them do work. So the idea of these bots being more companions Right. Like the movie Her. Have you watched that movie, JC, with Joaquin Phoenix?
Dr. JC Bonilla [00:09:15]:
No. I haven't, but I've kinda seen the the premise. Tell everyone what the movie is all about. It's a it's a really good analogy. What's what's happening here?
Ardis Kadiu [00:09:22]:
So that is a really good analogy. In the movie, Joaquin Phoenix has an assistant called Her, And this is an AI, and they form a bond and a relationship, and sometimes a romantic relationship as well. So it's the same thing here. Right? This AI assistance, they know us. They can anticipate. They can talk to us. They can be empathetic, they can connect in a much deeper, meaningful way when we think that generation z is kind of the lonely generation, and We talk about how even though they're the most connected generation, they're still the loneliest generation. These technologies have the ability to connect in meaningful ways, And they're gonna gravitate towards that, and that's why we're seeing those stats be so high for character AI upwards of a couple hours a day being used.
Ardis Kadiu [00:10:11]:
The larger players, like, for example, Meta with Instagram. They have introduced personas in there as well. So they're using celebrities in order to personalize different bots so you can chat with different characters in Instagram, I believe, as well. So this is something that is not going to go away, but but well, Let me take a step back. Let me ask you about the brands and how you think about like, how does this fit into a brand now? You no longer you have to have a website. You have to have your presence on Twitter, but now you have to have personas and brand characters in all these other places as well.
Dr. JC Bonilla [00:10:51]:
All marketers listening to this, you know that a brand has a voice, a tone, a personality, if you will. And if you're a financial person, you'll say also, hey. Brands have, you know, a commodity and a value. So, yes, all these things apply to a brand. I can speak to you in the brand tonality, the voice, and communicate yourself. Coca Cola. Imagine Super Bowl moments. Imagine Michael Jackson.
Dr. JC Bonilla [00:11:17]:
Well, Michael Jackson was Pepsi, I think. You get the point. Right? Or I can talk to you in a Mastercard moment, Visa moment, a Nike moment. Just do it and so on and so forth. So brands, the potential to learn how people communicate and they make a brand a lifestyle, Fascinating. Patagonia, a brand style about social benefiting, how, you know, as they sell things, they transfer that wealth to the world versus the brand that is a bank like Goldman Sachs. Right? So did aspect of how they think about sustainability? It's totally different. So you start seeing the brand emergence and, of course, you can totally capture it.
Dr. JC Bonilla [00:11:57]:
And you know what's really interesting on this? The way it relate to a brand as a Gen zer versus a millennial or my parents. Very different. And you're able to do that segmentation and targeting. So you can be the Coca Cola application for Gen Zers versus my parents. Right?
Ardis Kadiu [00:12:13]:
How does this translates to higher education? So as higher education institutions and leaders, They're looking at this technology, and they're saying, well, yeah, of course, I have my bot on my website, but I had to create all of my questions and answers. And and I know exactly what he or she or whatever gender the bot is or the name is. I know exactly what they're gonna respond because I had to painstakingly say, if the student asked this if the visitor asked this, respond with this. If they asked this, respond with that. So how should they think about it? You're saying that they need brand new bots that are able to handle anything?
Dr. JC Bonilla [00:12:48]:
Of course. In with an s pro. Going back to brand, I'm gonna take you back to, I think, 2015, and I'm doing this big, big branding exercise for the Graduate School of Engineering at NYU. And here's the brand tonality and the brand voice. You know who defined that? 7 tenure faculty members who really had a deep your knowledge about electrons, silicons, materials chemistry, and probably robotics. Right? I mean, that brand Persona tonality couldn't be more removed from how a 17 year old experiences NYU. Right? So what I'm talking about is that As we start thinking about these bots and applications, I, JC, 720 year old JC, can come and write a bot of what the NYU experience feels like for me. And then you start seeing that there's this, not brand dilution, but brand is interpreted by the consumer, by the user.
Dr. JC Bonilla [00:13:44]:
And what this is allowing me to do is literally discharge that onto the brand and still be the brand. I'm gonna superimpose three zero four tonalities. That's where my faculty members come in. But the 17, 18, 20,000 of the attributes that make the chatbot our design by little j c 17 year old with the aspirations of cafeteria talk, clubs, and I don't know sports. Those things were missed by the faculty members because it was all about research, no lariat of kinda academic awards, and that's important. It's just that we get so much depth in the way we get to segment or hyper segment. It's so promising and exciting.
Ardis Kadiu [00:14:23]:
Has to be personalized to the audience and to the context Of course. That you're in. So if you're a prospective student, you would click and you would chat with the faculty member that is the research expert. But if you wanna talk about admissions and admissions processes, you talk to the admissions adviser bot.
Dr. JC Bonilla [00:14:42]:
Well, I love that.
Ardis Kadiu [00:14:43]:
And if you wanna talk to someone who you want to understand as a peer and saying how are the dorms and how is New York City to live in. You wanna talk to a student bot or you're a student ambassador bot. So you don't talk to 1 person. You don't use 1 persona. You don't interact with 1 type of person at work or in the world. You're seeing the transplantation of the personas and and the types of work and the context that you're in Mhmm. Those being embodied in these bots. So depending on what you need and what context, you're gonna interact with a different bot.
Ardis Kadiu [00:15:19]:
We can take that even further and say, Well, what happens for current students then it's like each service or each office needs to have a different persona on that bot that is going to embody how you talk to somebody if you're in the tutoring center versus if you're in the Health Center or if you are the academic dean and you you're trying to appeal a reversal on your grade. So those different personas need to be thought about differently for each office.
Dr. JC Bonilla [00:15:55]:
What we're starting to unpack here, which I really love this conversation that we are organically expanding. Artisan and I did not know we're gonna get here, by the way. We're talking about the sizing of a custom bot versus the sizing of a chatbot as you know it today. Right. So there's 1 bot to rule them all, and it's a big investment. Today, you can have I'm not joking. You can have 252 bots that serve 252 different use cases. I'm gonna give you the most narrow specific example I can think about.
Dr. JC Bonilla [00:16:29]:
When I left academia full time, you know, I jumped into this kinda weird thing that I do today, wonderful in academia and then practice right in industry, what kicked my butt was Slack. I was not used to talking in Slack. Right? And all of a sudden, what I have is I know how to write an email, which is 257, you know, words, and it has a structure of a lot, and that's how I communicate. But now I need to dump 3 emojis and do it in a Slack way. So what I needed back, you know, 15 years ago, right, I needed my Slack chatbot, Right. Which is just tells me the things that I wanna say, how do I put them in a Slack? Now I've been failing and winning in this process, so now I don't need that bot. But nevertheless, that's the use case here that these custom bots can be so granular and support very narrow use cases, But that's the win, the scalability that they have. Back in the day when I wanted to launch a bot, it was about a year worth of a project.
Dr. JC Bonilla [00:17:25]:
It took a year. Never mind how much money it costs. Today, we're talking about maybe 3 minutes, and then a chatbot, custom bot, is ready to launch. And Maybe the GPT store will, you know, clone it and do many other things. So very interesting, but that granularity of the bot is where he's at.
Ardis Kadiu [00:17:44]:
Yeah. That's so that's the way we're thinking about it at Element for building. So we have our bold bot. Then we're building personas on top of it. So it has certain skills. It has different personalities depending on what type of context is it in, who is it talking to. Right? So it's the same technology. It's the same underlying information.
Ardis Kadiu [00:18:03]:
It's everything is the same. It's just that we can assign different personas in the context of where that is. And a persona is a combination of what the tone and the voice Right. And and how it's responding to you, what tools does it have access to, and, also, what types of information does it have access to. Right? So those are the things that change the persona. So as we move into the world of chat GPT, Everybody was learning prompt engineering, and that's something that we believe you still need to learn in 2024 In order to build some of these bots and custom bots, this week, OpenAI has launched the GPT store. So, essentially, their custom bots are called GPTs. I don't know why they picked that name.
Ardis Kadiu [00:18:52]:
It's very confusing. But, Essentially, GPTs are custom bots. We've been playing around with this since November. We've been building our own, anything from copywriting to hook generation to things like how do you do coding and and styles and so on and so forth. So we use those internally. We've shared a few of them, things like graduate growth hacker, you know, that knows all about graduate admissions and tactics and how to produce content that way. So now that the store is opening up, the ability to curate all of this in one place and monetize it becomes a reality. And what that means is that this is the largest platform or GPTs where a lot of people associate AI with chat GPT.
Ardis Kadiu [00:19:41]:
It's the most popular one. We didn't talk about Po.com, which is another platform that you can go if you don't like chat GPT. But, specifically, the GPT store of OpenAI is something that we should be paying attention to, and that's the reason why we're putting this episode out there, which we hope is very timely as this comes through. So we're gonna dive very quickly through some of the store features and kinda what they're releasing in the store. So, JC, as you look at this, what are those Different Components That Excite You.
Dr. JC Bonilla [00:20:12]:
So number 1, it's the monetization aspect and the aspect of creation too, right, and access it. So there's 3 things that I already know that need to be happening, and it's a in a way Mhmm. I'm just superimposing how my, you know, App Store looks like. That is, I can download as a user. I can upload it as the developer, and I'm gonna have some type of UX to the hospital Kind of getting in there, in there, and I'm expecting some type of, in that customization, ease of use, some type of no code, very intuitive QA in aspect, right, because, you know, if the chatbot doesn't work and things of that sort, what happens to interoperability Yes. That I did my chatbot in GPT 4, and you have 3.5. All those things need to be done at face value. 4th thing, which I'm very excited about and going back to I've done a GPT that it's for Indian students is that the discolourity and the diverse aspect of applications will have some type of tagging.
Dr. JC Bonilla [00:21:18]:
So if I'm in India Yes. India stuff surfaces. If I am in New York, US, New York stuff will surface, and then all that thematic jazz will take place. It'd be very interesting because I don't even know how you play a taxonomy there, so maybe it will learn. Maybe you actually do AI, generative AI on your generative AI GPTs, but I mean, I'm expecting that. And the 5th one, so you can come and break this all up and and elaborate. I hated this app. One star.
Dr. JC Bonilla [00:21:49]:
JC's ethos is the best. Five stars. And, goddamn it, the artist chatbot on how to make a mean Manhattan, which, by the way, this gentleman artist makes an incredible Manhattan. How do I download the recipe and the variations based on, I don't know, the type of alcohol that I have? Boom. So I dropped my comments, and then all of a sudden, it's complete. That's literally what we're expecting from looking at the App Store, how it works, and what at the minimum should be superimposed. What else will you add to those things?
Ardis Kadiu [00:22:21]:
I think you hit on all the points. You know? The the details are exactly what you're talking about. So the idea of having a store, it's a discoverability process. Us. Right now, there is a lot of sites that have popped up in order to aggregate different GPTs or different bots that can be useful. But the ability to have them all into 1 place, that's super important. If you're providing something of a lot of value and you're adding your data and you're adding your you need should be able to gain some royalties and to have some revenue out of that because of your expertise. So now this empowers people to look at this as I'm going to provide very high quality content and very high quality GPTs and prompting in there so I can sell this.
Ardis Kadiu [00:23:04]:
It's very similar to the free apps versus the paid apps. When you see a dollar sign on an app, You think, oh, this is quality versus when you see something that's free, you're like, I have no idea what the hell I'm getting in here, and probably they're gonna ask me for something when I get in there. So there is a signaling component of an App Store that the value of something that kinda correlates to how good it is and how much value you're gonna get out of it. So that's super important as well. Hey, Artis.
Dr. JC Bonilla [00:23:31]:
I have a question for you. How do you think this will play out? Many of the features in free apps today, It's that you don't pay because you observe commercials, these little banners that shows up. Basically, it's either free, paid, or that thing in the middle. Right? Which means that you don't pay for it as the user, but the developer gets something through advertisement. Do you think that the GPT store will have that middle ground of including advertisement in AVS. How does that take place? What do you think of that?
Ardis Kadiu [00:23:58]:
Yeah. It's gonna be really interesting. I don't see a a middle ground where you insert advertisement in here or third parties. That's certainly one of the areas where I don't know how we're gonna monetize that.
Dr. JC Bonilla [00:24:10]:
You You know what I just thought of? The GPT pop up. The GPT pop up. Oh my gosh. Can you imagine if it shows up?
Ardis Kadiu [00:24:17]:
Yeah. Every other sentence, it inserts an ad an ad read. I think that's a really interesting concept. It's like, how do you support these via ads. I don't know what the equivalent of that is. Alright. So as we're very excited about this, The store is opening up. You should go ahead and take a look at it as it becomes available to you.
Ardis Kadiu [00:24:40]:
You have to have chat GPT pro in order to access it. If you have not built GPTs yourself in the past, this has been available as functionality for since November so you can build your own. You could share them with links or you can use them by yourself. But now you're gonna have the store, so you can go and search for other GPTs that people have done without having to go to an external site. And, of course, hopefully, you get to c higher quality content and higher quality GPTs. And we highly recommend highly, highly recommend that you start thinking about building a GPT specific to your institution and your brand. So if you're, you know, NYU, You wanna build an NYU GPT for admissions and NYU GPT for alumni just so You can occupy that brand and that store location, so to speak, or that name on the store. So that's the first thing that I would do if I were you as a brand.
Ardis Kadiu [00:25:43]:
I would still go ahead and build that exactly like Jesse said for that advertising aspect. So your logo pops up when you search for your institution there specifically.
Dr. JC Bonilla [00:25:53]:
I just literally think of the one use case if I was back in higher education full time in my role of And, you know, academics, administration, that kind of hybrid. Here's the hack. I would grab the university catalog. I would LLM. I'll put an LM, and I'll create an interface, a bot, so that I can interact on how do I graduate. Remember how complicated it was? Hey. Am I gonna graduate? Literally, right now, you have the conversation, the interface that allows you to say yes or no, and what is path forward? What an amazing and so simple use case. 3 minutes in, university administrator, academic advisor, 3 minutes in, that should be your 1st use case.
Dr. JC Bonilla [00:26:39]:
Load in your university catalog and then let the student converse.
Ardis Kadiu [00:26:43]:
I do have to warn for somebody who's created a lot of them and looked at how you can upload content. Your content has to be AI ready.
Dr. JC Bonilla [00:26:52]:
What does that mean? I thought AI read anything.
Ardis Kadiu [00:26:55]:
There's a misconception about that.
Dr. JC Bonilla [00:26:57]:
Do tell us.
Ardis Kadiu [00:26:58]:
When you're uploading PDF documents, the traction of how that PDF is created. It's going to essentially extract the text from that document. So all it's doing is doing it's not doing an OCR necessarily, but It's doing a text extraction out of that. So there's a lot of libraries and techniques that we use internally at Element from extracting this information from different types of documents, website. But if you're just uploading a PDF document, you need to be careful about that because it doesn't read a lot of the things that you think it's going to read. So the best way to do it is to upload what are called CSV files or text files, just regular text files. Markdown files are also very, very useful as well. Markdown is just another format that you can specify, you know, headings and tables and so on and so forth in a textual way.
Ardis Kadiu [00:27:48]:
But they're very, very good for machines to read those because they are text files with just some extra, you know, components in it. So that's what we found to work really well with GPTs, building them over the past couple of months. So be careful. Don't upload PDFs. Try to upload AI already content. If you're uploading a catalog, it's better to upload a CSV that contains all of the headers and all the rows with courses and what they mean, then uploading your catalog, which is a PDF, which is a version of that catalog, but now it it's formatted depending on how that designer thought it was gonna be formatted underneath. So the text. Might be a little bit harder to extract.
Dr. JC Bonilla [00:28:28]:
Excellent point. So maybe you're not 3 minutes out from your chat GPT the custom bot on academic advising. Just need to put in 3 hours on getting the right format and then apply it 3 minutes. And to remind everybody, right, prompt engineering, data engineering remains a thing.
Ardis Kadiu [00:28:46]:
Yep.
Dr. JC Bonilla [00:28:46]:
And it's really, really important that all these skills permeate your campus. I've seen how universities and many other places just sometimes don't want to invest on the resources needed to prepare the data, and it's the thing. University officials. How many data engineers do you have in your staffs? Probably very, very seldom. Artist, have you seen a data engineer already sitting in a university?
Ardis Kadiu [00:29:08]:
There's a lot of chief data officers and so on so forth. I don't know if they call them data engineers in universities, but, certainly, they kinda sit in the IT departments. So, yes, to your point, having AI ready data is super important. Text, video images. They're all now considered data. As this model's become multimodal, that's something that we have to all be cognizant about. So the first thing about having our AI ready campuses, it's going to be how do we look at our content, how do we look at our data and make sure that it's available for all of these large language models to work with that particular data set. And as we move into the custom bots world, think about use cases and think about workflows on how can you package a whole bunch of content, tasks, processes, and they can your custom bots, and they can become those packages and those apps that you can now distribute to your teams.
Ardis Kadiu [00:30:12]:
You can have your teams be empowered and produce the same amount of work. So that's how we see work being packaged rather than specific features of functionality. It's gonna be around packaging workflows with content, with prompting in order to to kinda get to something that's a custom bot. So with that, I think we're going to wrap this one up. We've talked and we've gone in a lot of different directions. Hopefully, this is exciting and and everybody starts playing around with GPTs. We're certainly gonna put a whole bunch of them out there.
Dr. JC Bonilla [00:30:45]:
JC Ethos, look for it.
Ardis Kadiu [00:30:47]:
Appreciate it. Thank you, JC. Till next time.
Dr. JC Bonilla [00:30:50]:
Thank you, everybody.
Ardis Kadiu [00:30:53]:
If you guys like the podcast, please subscribe to it on Apple or Spotify or wherever whoever else you listen to podcast. Rate us 5 stars. We are sometimes entertaining, sometimes not, but definitely, we're gonna pack a lot of AI for you. I'm telling you what that means. Thank you.