Revolutionizing Training with AI: Insights from Luis Garcia

Brent Peterson (00:01.711)
Welcome to this episode of Uncharted Entrepreneurship. Today I have Luis Garcia. He is the president of PEAT. Luis, go ahead, introduce yourself, tell us a little bit about your day -to -day role, and tell us one of your passions in life.

Luis Garcia (00:16.711)
Okay, well, thank you, Brent, for having me. So, yes, I'm president of Pete, and Pete is an AI -powered learning platform. But what that means is that we provide tools for growing organizations to create training programs. And small organizations, medium -sized organizations, let's say the low hundreds, usually when they're growing, they start facing the problem that you no longer can shadow to learn for new employees.

for product knowledge and things like that. And then they start relying on their experts and functional managers and leaders to become trainers. And they realize that it's a lot harder than they thought. So we're creating training for content, content for training. The process of translating expertise into training is complex.

And usually you have to pay people to do that, have training departments. Smaller companies can really do that. So our software can do that 20 times faster than the traditional methods. And so a company that doesn't have a training department can have one out of the box, so to speak, from our software. my personal journey, I came from the world of technology and education. And I started my career as an engineer.

and I fell off into a higher education by almost by accident in the early 2000s when there was a one period that only people that older like me remember that one time that technology actually was went through a slow period. And so I joined a university and they wanted to create an online university and because it was a traditional campus university and they hired me to do that because I knew how to make websites. And I ended up.

been a very great journey. So I've been in educational technology for 20 years. I built a very large, fast growing online university that ended up today has 18 ,000 students. And in that process, I learned the challenges of creating content for training. Because when you come from traditional education, if people are sitting in the classroom and the educator is a good communicator, he engages way with you.

Luis Garcia (02:40.571)
then people feel that they're being educated. But when you remove that and now you've got to go online and get materials from a website called Learning Management System, then these materials really matter. So I have to create a department to create the materials for translating the courses from campus to online and all that. So I'm very familiar with this problem. And the other problem I'm really familiar with is the evaluation of learning, what we call assessment.

And how do you know that the person actually learned? And so we use tests traditionally, and that is actually a very bad way to assess learning. And the best way is to really personalize the engagement with the learner and ask questions, inquire, and go through a train of thoughts, simulate situations. A test is the worst situation because it doesn't exist in the real world.

and you don't go into a room, don't ask anyone, can only answer in two hours, that is a very artificial situation. So it's better to do actual simulations of real world. so that's the other problem we wanted to solve at Pete. So we have tools that simulate situations and simulating structure as well to ask questions personalized to each learner and to do learning assessment at scale. So content development and assessment of learning is what we do.

Brent Peterson (04:05.442)
Wow, that's really good and I like the way that you approach that. How about passions? Do you have a passion outside of what you do for work?

Luis Garcia (04:12.231)
yeah. My biggest passion is my family. I have two kids and they're getting older now. I have one kid in college and the other one about to go. And I've been married for 28 years and so I put a lot of value in my family. But apart to that, my two passions are sports and music. And I run, I swim, and I play guitar.

Brent Peterson (04:37.126)
awesome. Yeah. So we're very similar. I was a programmer in my past life. I run, I swim, I bike. I play piano. I don't play guitar. we're pretty close.

Luis Garcia (04:42.747)
There you go!

Luis Garcia (04:48.815)
I used to play piano. don't play piano anymore, but yeah, fantastic. We're so similar in that. you do like open water swim? Yeah, I love that too.

Brent Peterson (04:53.588)
Yeah, funny. Yeah, mainly open water. we, yeah, I've done last year I did Pointe de La Pointe, which is from Bayfield, Wisconsin to Madeline Island. So it's across Lake Superior. That was my most scary swim, but we winter in Hawaii, we winter in Kona and I swim in the ocean every day there. It's fantastic. And then we summer. it two and a half miles.

Luis Garcia (05:16.919)
What was that distance that you did in Wisconsin?

Brent Peterson (05:23.438)
And I'm a terrible swimmer. My daughter swam with me. She finished in an hour and I finished in an hour and 40 minutes.

Luis Garcia (05:29.863)
So I, in my forties, I did 10 Ks, so I six miles, races. And, but my most memorable is the Bay of San Francisco, which it was only like a mile or so from the island, where I could trust into the beach. This should do that one. It's pretty amazing to swim that bay.

Brent Peterson (05:35.042)
That's crazy.

Brent Peterson (05:50.606)
Yeah, that's definitely on my bucket list. That's escape from Alcatraz. Yeah. Good. Well, we can talk about that offline. We don't want to bore everybody with sports talks, but thank you. Before we get started though, I'm going to tell you a joke and it's guaranteed to be a very poor joke. All you have to do is give me a rating one through five. So here we get, it's called the free joke project. So here we go. Socks should be at least 12 inches long so they can cover a foot.

Luis Garcia (05:54.085)
Yes.

Luis Garcia (05:58.757)
For sure.

Luis Garcia (06:23.025)
That's clever! I like it. I give it a three.

Brent Peterson (06:24.782)
Okay, thank you. I would have given it a two, but I appreciate you being so generous. All right, so I love the journey. I like what you said about taking a test because I used to be in the Magento community, which is a software platform, and I always participated in the educational process in terms of being a subject matter expert. And they went from just doing an A, B, C, D type of test.

to be certified on Magento to more of an interactive test where you had to give answers and give code related problems or give answers to code related problems. And I know the goal was always to make it non -memorizable, right? And I think what you're kind of keying in on and unfortunately you only had two hours to complete the test so it wasn't real, but I know they had to figure out some way to do a certification.

I kind of want to go backwards from your story and start with that part of it on how you work with students to, or even not students, but how you work with businesses to ensure that they know what they've learned.

Luis Garcia (07:34.491)
Yeah, so because I come from that world, I know there's a very difficult problem to solve a scale. so a better way of assessment could be project -based learning. instead of asking people to answer a test, you ask them to do a project. that's not something that you memorize and you exercise. It's a better way of assessment. But it's also very difficult to scale because you need an expert on the other side to kind of assess the.

the project. So it's a very complicated, a very complicated problem usually in education we have in that time is fixed. You know how long people are sitting down there or taking the course, but learning is variable. So the two different people in the same classroom, virtual or physical, you know that how long they were staying there, but they could have learned, mastered the material of much different levels. So in the education has been these nagging questions for

for decades now and say, how do we make it the other way around? How do we make the learning to be fixed so everybody masters the material and the time variable? And so some people may take less and some people may take more. And again, the only way to do that is to have enough experts to engage and spend enough time with each learner to do that. And it's just not scalable, not even in the classroom.

But with AI, you can. And you can train an AI to be the expert and then have a conversation with the learner. they can do that hundreds of conversations at the same time, because it's a machine doing this, No limitation on that end. But with that same knowledge that you have, and by the way, we get this knowledge because we created the materials to make the course. So we have a knowledge base.

that is created by the company. so I tell you a couple of use cases. We have a retailer that has a social component to what they do. They sell liquor, they say wine. And so when people buy from them, they can choose a cost that they want part of their purchase to culture. And so it's very rooted in their culture to do this when we're

Luis Garcia (09:56.515)
So when they bring new associates, they want to make sure that that's at the front of the every customer interaction. So we programmed the simulation engine to come and train the associate is, okay, here comes Luis. He bought this wine last time and he's interested in supporting our troops. So you have to welcome that person with, Luis, welcome back, you know, ever since last time.

your contribution when to do this and what do we want to do today. So do you want to continue with your purchase today to support this cause or you want to support another one, animal shelter, whatever it is they have. So training that and with an associate will require a senior associate to be with them one -on -one doing that kind of training. So we can simulate that situation and then we'll give them the transcripts to the leaders if they want to do more one -on -one follow -up but at least they don't have to do that.

So we have a bank that is using our simulation engine to train the tellers and with all the different situations that can come up in on the window and Some of them are very usual but then there's that five percent that is very unusual that they have to start calling people to ask we can get those conversations in there technical teams are using it to train on tech stack and you welcome a new programmer and I

You're not going to teach them how to program, but there are things the way you do it in your team that are special to how you do depending on what your tech stack is. So we can, instead of putting them to do a test, they can still do the test, but do it personalized based on their role and what you need them to

Brent Peterson (11:39.958)
That's good. think in terms of AI, like what you said earlier about just about the technical learning that seems to be one of the best applications or one of the biggest applications for AI besides generative AI. the idea that you can have something assist you in in helping your job go better, let's just say that's a great application of how that goes. Do you find that

that you get a little resistance from some clients who don't quite trust the AI yet.

Luis Garcia (12:13.613)
No, if they don't trust AI, they're probably not our customers. so the customers that come to us are very eager to bring AI into the training. And they have seen how we can assist them in other parts of organizations. sometimes they only have a training problem, and we just happen to be the solution. And AI is the power of that, so they don't really mind.

But we have seen some customers that they are skeptical of AI, they want to bring it into the organization and they don't become customers. But it is a problem because it will happen no matter what. It will happen with us or with somebody else, but AI is going to your organization whether you like

Brent Peterson (12:59.308)
Yeah, that's very true. One issue that I've seen that others have talked about in terms of using private AIs or an AI to train is right now the acceleration of what's happening within the AI space is so rapid that one version to the next often breaks your training material that's going into the AI. And because that trajectory is so fast going up,

it's sometimes harder to keep up and do you struggle with wanting to go with the latest but then kind of holding back and staying just say with whatever GPT -4 and when you could now go to this zero one preview things like that.

Luis Garcia (13:40.209)
Yes.

Luis Garcia (13:44.251)
That's a really good question. so when people are doing it themselves, you can today write a custom GPT for yourself and you're writing. will. And that custom GPT, first of all, it's getting your information and making it public. So that's a problem. unless you have a private one. And I like to say to our customers, we take that problem away from you because you're not doing the one doing the custom GPT. Our platform is.

agnostic to the LLM. We'll try to make it as agnostic as possible from the LLM. we're orchestrating. We have our own AI that is trained as an instructional designer. Instructional designer is the person that knows how to translate knowledge into training materials. so what we do is try to find the right algorithm for the right task. And in that journey sometimes,

the algorithm change and evolves. so we take care of that one particular part, but we're not using one algorithm for everything. So we use different depending on the task that we're trying.

Brent Peterson (14:51.084)
Yeah, I think from a business flow standpoint, that is an underpinning that people don't understand, especially when you're trying to do it at scale. If you're going to use llama, there's a new version that's for a hundred, four or five billion parameters or something like that. There's all kinds of things that people don't understand from the backend and that you're constantly iterating and making it better. Let's talk about your business though, because I do have a friend who does.

Luis Garcia (15:06.161)
this.

Luis Garcia (15:14.652)
Yes.

Brent Peterson (15:18.51)
this for a living but with real humans, right? And you know, they have a $6 million business for just training for exactly what you do. Is there a threat to that type of business where they are doing that custom training for larger corporations, where maybe they wouldn't need that trainer anymore?

Luis Garcia (15:37.895)
I will say this, today we target companies between 50 to 500 precisely because it's expensive. So in the market, only companies over a thousand employees will have the resources to either create their learning department or hire a company like your friend and do that. But a company that is 200 people don't have half a million dollars a year to do that. They may have it, but they're not going to apply it to this. So that's why...

we target that particular market. We don't think anybody's really targeting them and try to solve that solution because it's expensive to do courses. And the closest it gets if you use one of the biggest providers like Coursera or something like that, and there's relatively inexpensive to license from them. But you can get an onboarding course of your company from Coursera. You can't do that. It's not customized to you. So that's what we can do. And the two customizations of courses are very low.

at a very low cost. So to answer your question, I don't think it's a threat today because we're not really trying to unseat. The market is very, very large in the target that we want to go after. And if I was those companies, I would be using AI to increase my margins. And I do think the threat will be in the future that people that hire those companies are going to realize that they can get the same, and it's not the same today.

the quality of production value, if you will, is not the same. So they still have that edge. But that will be reduced over time and maybe their prices are gonna start coming down. Fully replacing? I don't know, it's hard to tell right now.

Brent Peterson (17:19.01)
Yeah, I think so. You know, my, have a new company that we, we do generative AI, but we have humans that edit it, right? We have sort of people that always want to look for it. And I always say that when, when you put something for sale, Google is sell is helping you get somebody to buy it, but chat GPT is never going to buy it. Right. So at the end of the day, a human, and in your case, at the end of the day, a human has to get trained. how much.

I guess my question now is how much of the human interaction is involved in your process to ensure that it's always correct?

Luis Garcia (17:51.387)
Actually, I'm so glad you bring that up a lot. Actually, we have to rely on the expert to curate, to give us the direction, the vision. In a world where anybody can create content, what really matters is the point of view of the expert. And so we try to capture that as much as possible. So nothing goes directly to a learner without having been to an approval process from an administrator that usually is an expert in what we're trying to create.

So they have the power of doing that. We also capture their voice because that's part of the personalization. Brent is the expert in that. And right now we don't have to put him on the microphone for hours and hours and hours doing takes on the content of a course. We'll clone your voice and then when we generate those courses automatically, in an automated process I will say that instead of automatically.

and we'll use your voice for that. So the learner will hear from you. But nothing goes to a learner without having been approved by an expert first. So we're really just assisting the expert, not replacing it.

Brent Peterson (19:01.356)
Yeah, I think a lot of times people think that AI is going to do everything for them from end to end and they forget the step of, is this actually readable? no matter what you're going to do, it's never going to be. I mean, I guess, again, we go back to the point that this is meant for humans to, in your case, humans to learn. And we need to make sure that it's learnable and that you need to learn it, right? It's relevant to who it is. Subject -granted expert is such an important part of that.

The other part is, I think that people do make the mistake and I know that there's LMS and I think you're not, would you call it -first as a learning management platform, LMS?

Luis Garcia (19:40.315)
We have a learning management system as part of our solution, but I don't call us an LMS company because it's not our focus. But the courses have to go somewhere for people to take them and they do land in an LMS.

Brent Peterson (19:43.469)
Yeah.

Brent Peterson (19:46.816)
Okay, well, yeah, yeah.

Brent Peterson (19:53.718)
Yeah, I think if you let's just, let's just say that there's somebody that has sort of this sort of all AI LMS and they make the mistake of just putting it out there as a done course, which in, think in generative AI right now, writing a blog posts, so many people make the mistake of copy and paste from JCPT and send without editing it. I do that myself often when I'm

looking for a social post and I think, and it's only 140 characters, who cares, right? And then you realize you've said something really silly. From a professional standpoint, that's worse, right? You're now putting your reputation out there and it just goes back to what you said earlier, having that subject matter expert. going back even, let's take a step back then from your journey. How did you decide to start?

Or how did you decide to get involved with Pete and how did that journey happen?

Luis Garcia (20:50.865)
So it's a good story. I have a co -founder, and he is a very well -known tech entrepreneur here in the Orlando area. He built several companies, exited them. And his last company, went Unicorn, very successful fintech company here. But in his last couple of years in that company, they acquired other companies.

and that they needed to integrate all these new people into their organization. And he took it upon himself to create a center of excellence and do this thing, you know, create courses and all that. And he realized how hard it was and why he thought it was gonna take him a couple of months. It took him six, eight to eight months. He had to hire instructional designers. He said it was thousands of man hours. And so when he exited...

his company, he said, this is a very interesting problem to solve. We have met and we have been friends from the local tech ecosystem here. And he knew that was, you know, a higher education. had working at tech for a very long time. So we went out to lunch and he tells me, you know, I'm thinking about doing this. you, would you help me with that? And I was like, yeah, I love to. I think that it will, it will be really great to use AI to solve some of these problems that I know exist.

And so we joined forces last summer and we got started.

Brent Peterson (22:18.318)
That's awesome. I have a question just about cultural aspects. In my previous life, I ran a company where we had office in Mexico City, have offices in Bolivia, and the culture between all those different areas was so different. Do you find it difficult to sort of engage in the Latin American culture compared to the US culture, and I'm assuming you're probably in Europe as well, just making those learning, making it

work in all the different areas of the world or right now are you just concentrating on the United States?

Luis Garcia (22:51.911)
So most of our customers are in the US. We do have a customer in Latin America. So was the first time that we were actually using a platform to develop courses in Spanish. And I have a lot of experience in multilanguage development. That's another difficult problem to solve. And usually for you to have a platform that's multi -platform, multi -language, that has to be thought out for the

very beginning. It's really hard to say if you have a SaaS platform and say, now I'm going to make it in Spanish, you realize if you haven't thought about that from the very beginning, that's a very difficult thing to do. But because we use large language models, in theory, they can create in several languages. But we never had actually tried it. And we got a customer from Latin America wanting to do it in Spanish. And all those questions are starting to arise. OK, so we're going to.

have the voice. Okay, is that going to be with a Mexican accent or is it going to be for an Argentinian accent or is it going to be a neutral accent? And so the person provided us the the the cloning of voice and then there was a lot of challenges into finding the right tone and and and all that. it was not not to mention the QA portion. We to then get hired somebody in Latin America to do the QA of the courses.

And beyond the curation of the courses, just making sure that every word is pronounced properly and the output was generic enough for all Latin America. So yes, it is a challenge. And I worked in different countries before, and so I know about some of this. they continue to apply when you put technology in front of them. AI is not going to make it easier.

What makes it easier is that we're still creating content faster than we would otherwise. And that's great. But there was a lot more nuances to it.

Brent Peterson (24:50.218)
I want to talk about some specific applications and, and I also saw that you have a fractional business where you help fractional. I just read about it on your LinkedIn, but when I exited the company that I was at, I took on a role as a fractional CTO and I got to learn more about private equity and how that works. And then in larger scale, what if I were to say to this company as an advisor, Hey,

Luis Garcia (24:59.483)
Yes.

Luis Garcia (25:06.598)
Mm

Brent Peterson (25:17.774)
that you could use this process, and they have 200 employees, if you could use this process for your onboarding, for sales training, is there specific areas where you see, let's say, low -hanging fruit in companies that don't have anything implemented in terms of any internal training?

Luis Garcia (25:36.027)
Yeah. experience tells me that usually when you start a training department and the first place that they go is on board. Because you want to, there's no training department in the world that feels like they are doing their best or they're maximizing because there's always more needs. And so you always want to say, well, what can I do that will touch everybody?

onboarding it. So that's usually the first thing that a learning department starts. And then the second thing that they usually do is sales training or product training because that's the thing, you know, is very, very important for the business. So I thought that those were going to be the use cases that we will have to work on, but they have not been. And we have been approached for very niche cases in...

Earlier in our journey, we signed on with sports teams and sports venues here in the Orlando area. And their problem was that they have 50 or 60 employees that work at the venue, but on game day or event day, they have 200 of people that come in only to work on the event. And training those people was very challenging.

and because you can now all get them at the same time. If you do it at the beginning of the season, they're likely change over the season. really they kind of live with this problem that, okay, I hope it goes well. And creating a training department for that is not cost effective because of the amount of retraining you will have. So our tool fit perfectly into that particular problem because we can create a content fast, we can get it to everyone.

in that we can retrain very easily. And that's a problem we never knew that we will have to solve. And then we kept signing more venues because they had the same problem. And it's very customized to them. And so the head of securities and safety will say, they don't know anything about safety and security. I have all my processes here and all I can do is give it to them and hope they read it. And I mark them in the list like, I give it to rent.

Luis Garcia (27:54.171)
So we took that manual, turned him into a course, synthesized his voice, and now he's the voice. They hear directly from the head of security and safety in a course for their level and the things that they need to know. then now only can Mark them in the list that they got the process of the Mark them in the list that they actually took a course and learned. so that was a use case that we were not expecting at all.

Right now we're just signing a university. We have not targeted universities for our tool because higher education has particular needs that we're really for corporate training, but universities do have staff training that you have to have. You have to train your people, especially in the areas of financial aid and admissions. Those are complicated jobs. I know that firsthand because I work on that, and that problem myself.

when you hire a financial advisor, it takes months of training before that person is capable of helping a potential student with a financial aid packaging. So I knew if back then when I was growing the online university, I would have a tool like Pete, I would have cut that time in half or even 75%. And that's real money of resources that when you're paying a salary where the person is being trained.

a higher education trade school recently approached us and I said, yeah, I know exactly your problem and we offer them and they're gonna use it for that kind of thing. So it's not really onboarding new employees, it's regulation training. so we just never know how people are use your tool.

Brent Peterson (29:35.052)
Yeah, that's good. Where do you see this going in the next five years? Where do see your industry going to?

Luis Garcia (29:42.331)
What is exciting is that I think that in our target market, training is something or lack of training or bad training is something that companies just live with. They don't think they can solve them. They just keep stumbling on it. And some people survive and some people don't. I'm just excited to provide a tool that can help them and grow. If that means that some of them become very large organizations and we grow with them and we end up being an ICP of thousands of employees.

then great, but I think what's exciting right now is that there are hundreds of thousands of companies that can use our product.

Brent Peterson (30:19.562)
Luis, as I - we have a few minutes left as I close out, I give everybody a chance to do a shameless plug about anything they'd like. But, you know, before I do that, I forgot to ask you, where did the name Pete come from? I'm interested in that. What does that mean?

Luis Garcia (30:34.075)
Well, actually, P stands for, it's behind me, in personalized education, training, and enrichment. And we wanted to have a four letter word that will be memorable and will stand for our values, which is start with personalized education and training. And so that's where it came from. And with that, with my shameless plug, then I go directly into that.

Brent Peterson (30:58.476)
That's good.

Luis Garcia (31:03.429)
We acquired the pete .com domain, so it's very easy for people to go. So it's pete .com. And you go there, and you can read about us, and you can request a demo. And we'll happily go and do a demo for you. And hopefully, we can help you with your training.

Brent Peterson (31:20.814)
That's great, yeah, and I've been on the website. You've got a nice intro video or a video that kind of explains it, and there's a lot of FAQs and lots of information. And I'll make sure I get those into the show notes. Luis Garcia, it's been such a great conversation and very interesting. Thank you so much for being here today.

Luis Garcia (31:30.118)
Yes.

Thank you, I appreciate that.

Luis Garcia (31:39.932)
Thank you.