We had the privilege of interviewing Peter Bergin, the Chief Product Officer and founder of Golf.AI. They started with a vision to transform the golfing experience through artificial intelligence, offering solutions designed to help players enhance their game. Over time, the company has expanded its offerings, continuously pushing the boundaries of sports technology. With even more innovative products on the horizon, Golf.AI is at the forefront of shaping the future of the sport.
In our conversation, Peter shared the story behind Golf.AI, the challenges of navigating the Sportainment sector, and his entrepreneurial insights. Whether you’re passionate about sports technologies or an entrepreneur seeking inspiration, Peter provides valuable perspectives for all.
Check out the full interview below or listen on YouTube!
Interview with Peter Bergin: Transcript
Topgolf.Club: So welcome to the Topgolf Club podcast, the show where we explore the business of sports, entertainment, and technology. I’m your host, Burak, and today we are diving into the fascinating world of AI in sports.
Joining us today is Peter Bergin, the founder and the chief product officer of Golf AI. A company that’s revolutionizing golf with the power of AI. Peter is here to share the story behind Golf AI, the challenges and what kind of challenges creating a sports AI solution is and his vision for the future. So hello, Peter. Thank you very much for joining us today.
Peter Bergin: Awesome. Thanks for having me. Excited.
Topgolf.Club: Yeah. It’s an honor. And let’s dive in. First, I wanted to ask about the background a little bit. Golf AI, what inspired you to create Golf AI? Was there a specific gap in the market that you aimed to address or how the story started?
Peter Bergin: Yes. We can go pretty far back actually. The story started probably when I moved to Australia before we were recording talking about it. Really two years ago and I actually started to get a lot more technical learning to code, learning to build applications, web applications, backend technology, various different things. And in that time is also when ChatGPT came out. So the whole AI excitement started happening. Obviously, AI has been being built for years and years and years.
Tens of years or hundreds of years, probably people haven’t been thinking about it yet. But that’s when I started getting very technical and being able to build these applications. And with that, as you start building technology, you tinker and you build different versions and allocations and stuff like that to test and build your skills. And me being enjoying sports and liking sports and having that be one of my passions and then also this AI movement was happening and me also trying to build the, build my technical skills. That’s essentially what I would do. I would build on these AI applications. And obviously as that industry progressed, I was able to carve out a little niche for myself, being able to provide some assistance or consultancy or contract work to do various entities.
So that kind of progressed through the last couple of years and then just through various meetings of individuals, we got together, there’s five individuals to start Golf AI. And the big thesis around that or the general thesis was that these language models will be obviously very important in our lives moving forward. Sports is still maybe one of the last live art forms, like maybe concerts and sports are very much the very last live art forms that will not go away. People like community, people like competition, people like people. So we like enjoying what we watch other people play sports.
So we wanted to play in that space where we knew both were gonna be very big obviously moving forward. And if we can play in that space, then we can figure that out. We spent a long time trying to figure out what we wanted to do in that space. There was a lot of projects going on that were doing a lot of R and D within that space. But specifically for golf, what we’re able to do is because golfers use technology and invest in technology and invest in their own game personally with both time, money, attention. Which is much different than any other sport. If you play football or soccer, generally you need a ball, shoes. You don’t really need an application to do anything with. And that can go across pretty much every other sport. So golf is the one where people use technology pretty conclusively in their own hobby or activity. And then also they watch the sports well because obviously what we can do with all these language models is go into the world of how you watch sports.
So we can both play in that cross section of while you watch and while you play. Which was the general thesis of Golf AI. There’s five hosts in terms of the partners at first, we’re just generally very connected in the golf space. There’s people that have started multiple golf companies, people that have played high level golf across like the PGA. Just the passion for golf, passion for sports and passion for technology and trying to figure out that space. And that’s the initial idea for Golf AI came through a natural progression of exploration and figuring out what we might have the best chance of succeeding at.
Topgolf.Club: Yeah. Sounds good. I also saw some AI solutions in golf and in tennis space that you can download as an application and make the phone camera record you and then it gives you some kind of tips like how you swing better and that kind of game improving.
AI usage and also some application usage was involved. So I definitely think there is a need, like you said, the golf players tend to spend time and money to get better at the game of golf. It doesn’t matter what level you play. And again, like you said, it is a different than a soccer player. I’m pretty sure there might be some app that you can use to look how to shoot the soccer ball better, but I don’t know how many people would be interested in doing that. Golf is obviously in a much better position with the data points and the whole usage of the new technologies and also some companies are investing a lot inside of it. So I think it’s spot on for Golf.AI.
We can deep dive with technology and innovation. After I knew that I was going to have an interview with you, I played around at Golf.AI a little bit. When I first signed up, it asked me to choose a partner. And I chose Hiro Nakamura because I thought the Zen Golf Master was a cool skill. For the listeners besides my partner, you can also check out the golfing rules with Golf.AI, which I think is an interesting point because in contact sports, it’s hard to ask AI, was it a foul or a penalty, but in golf, there might be specific situations where you can directly ask and find out about the rule.
So can you explain what users can do with Golf.AI? What makes it possible? Also, I saw there is something called My Bag in the sections. What will the My Bag section look like in the future?
Peter Bergin: Right now, we’re very much in the infancy of building Golf AI and coming out with a couple very exciting features hopefully over the next two weeks. One of the big first steps was solving the rules. The rule handbook is impossible to navigate as a normal human being. So we rely on other human beings on the golf course that are also bad at golf. With language models, we built this data set to prove the golf rules scenario. You can ask a question and get a very accurate answer. We tend to be accurate on basic questions. Golf rules can get ridiculously confusing.
Sometimes even human beings have a hard time. Right now, you can ask simple questions through text or voice. You can talk to your partners; your partner is the personality and voice behind the text or voice output. The other side is understanding the live golf landscape across the pro tour. You can ask real-time questions like how Collin Morikawa did last week. Along with that, we have a trending section on the homepage. Trending topics in golf. Click it and get an overview. What we’re moving towards is an on-course caddy. Your partner can be a caddy, understand the live golf landscape, know the rules. The on-course caddy will be an application that talks to you. It’ll pick up your location, understand wind direction, bunkers. We can do what a caddy would do standing beside you.
For My Bag, upload a video of your golf bag, and AI will understand what’s in it. We can analyze your bag based on your game, suggest upgrades, compare it to pros, value it. It’s a culmination of different areas of golf. The easiest way to say it is having a caddy in your pocket. I was testing it yesterday on the golf course. It’s exciting to have your partner talk to you as you go. We can qualitatively understand your position, bunkers, hidden objects. Instead of a rangefinder or another app, we can talk to you as you play. It’s a holistic platform for while you watch and play. That includes rules, local rules based on your course, pro game insights, and stories.
Topgolf.Club: Okay, great. So, well, you mentioned like it’s a cross-platform for news rules and in the future it’s going to be maybe your personal golf assistant that will make even recommendations to you. And when you were explaining, I also thought like maybe, if a person is playing at the same golf course all the time that with the data points gathered by Golf AI, it might also make some accurate recommendations in the future with the gathered information since that you will have some data set that makes sense to analyze with AI and create recommendations out of it.
Peter Bergin: Exactly. Yeah. So as you use the platform, it improves as we go, right? So which is exciting. It understands you a little bit better, understands your tendencies, and obviously with the data points we can use as we move around the golf course. I wish it was exciting. Another question is what were the biggest technical challenges you faced during the development of Golf AI? Was there something specific that you would say this was the real pain point when I was developing this?
Peter Bergin: Yeah, I mean, I think in the sense of using these language models, so we all use ChatGPT, or have, maybe some people have, some people haven’t. I’d imagine most people have at least seen it. So these language models are a lot of the backbone of how things work in the application. And these language models, unlike other applications, right? So if you’re using an actual application, you press a button. We know you press a button. We know where you want to go with that button. With, for example, a question you might ask on Golf AI, we have to do some searching in the background. We have to assert information. So if you ask a question asking about golf rules, your question won’t be similar to anyone else’s question. So we then have to pass it through these language models to figure out what you mean and what you’re asking. And we have to do a lot of, let’s say, let’s call it knowledge work. Automate the knowledge work that has to happen in the back. So therefore these language models are, let’s call it non-deterministic. We can’t exactly predict the output every single time. You don’t know for sure what it’s going to do every single time. And that’s the magic of it, right? They’re creative. They come up with unique answers and they come up with different answers. So there’s a little bit of a challenge in making sure we build enough guardrails in it that it is outputting the correct information. Like for example, on the golf course, if you’re on the golf course and we have these language models that are speaking to you, if it’s 120 yards to the bunker, it should say it’s 120 yards to the bunker. That’s, we don’t want that to be different. But maybe we want the way we say it’s 120 yards to the bunker. Maybe it’s in a different language, maybe it’s in a different tone, a different style, right? Like all these things. So you want the creative elements, but you also want to have these guardrails to be able to actually produce the outputs that also are relatively expected because they can go off the rails sometimes. So the difficult part is building in these language models to actually be predictable, but also within the scope of allowing the user to kind of like build their own journey as well at the same time. So it’s really around the language models, a lot of how it works is, so what you’ll see is you’ll type in a text box and the text will come out. Within that, that is one text input and one text output. But behind that is probably four or five other text inputs and outputs that are happening to run the system. So there’s a lot of capabilities for errors behind that. So just being able to manage that, handle that. And then also, as you just kind of build a system, you expect users to use it one way and then they come in and use it another way, right? And therefore you have to then build the system to be able to fulfill that capability. So, what they might think or want to use, right? It’s working alongside people that can give you feedback and obviously, and then be able to show you what they’re using it for. So it’s, but it’s really around managing the language models. And they’re getting much better. They’re getting much smarter. We’ll be able to fine-tune them to be able to be much more useful within our application and stuff like that. So it’s just managing that. But if you can figure that out, then it is quite impressive what they can do. So, which is exciting.
Topgolf.Club: Yeah, obviously. I mean, when I was playing around with the app a little bit, I also realized that it’s giving some really accurate answers. And I mean, I’m also someone who uses ChatGPT daily and other AI solutions, so I can say that was a really successful answer. I mean, I didn’t talk with Golf AI a lot, but like in fifteen twenty minutes, I was pretty satisfied with it. So congratulations, I guess, on the technology. Let’s deep dive into business and marketing strategy a little bit. First of all, I want to ask you like how large is your team? How many people are you and are you guys hiring more people? How is it working?
Peter Bergin: Yeah, so right now the team full-time is five people. And then, if we talk about like subcontractors, part-time people, that could expand to nine, ten people on any people that are working on Golf AI and stuff like that in various different capacities, some as needed, some hourly. So yeah, always looking obviously to bring out talented members. The whole idea of business, it’s essentially just a collection of people doing things that also provide value. And then the big side of that is, like you said, like the marketing user acquisition side, because we’ve been pretty deep building the technology and obviously now it’s kind of like this pathway to actually getting out to the world. In terms of marketing and distribution and stuff, so we have a couple really good industry partners that will be coming out with a couple features and activations with, let’s call it around the My Bag coming soon that you saw. There’s actually something in the back end that’s happening around an industry partner, which is exciting. So obviously, allowing those pathways for user acquisition is good. And then, the secret strategy, I don’t know how much is the secret strategy, is around the AI partner. So when you first signed up, you used the AI partner or selected a person. The trick is that there’s, if you scroll through the list, there’s some, there’s one of them that actually has a star on it. We don’t do a good job at showing what this is. That’s actually a real human being. Right. So that’s Ewan. Ewan is a friend of ours here in Australia. He’s actually a broadcaster for Sky, BBC, I believe. On the tour, so he goes to tour events, DP World Tour events, so like that to actually do commentating. He played a couple of professional PGA events. So the whole idea is then to be able to license IP and brands and human beings to be your AI partners, right? So obviously you can imagine like the players and LPGA players and different influencers and people in the industry that could be your caddy on the golf course. Won’t go too deep into promising any because I tend to promise the big ones, but we’re not supposed to do the big ones first, but that’s the whole idea, right? So it is one of them can be the caddy on the golf course, golfing alongside you and stuff like that. So obviously very similar to how we see it as. Like, when you play a video game, you play with the top players, right? Because I think that’s… In those video games, a lot of the times is if those players and their names and their image weren’t in that video game, the video game would lose a lot of value, right? So, so that’s kind of another idea we’re moving towards. Which I think is exciting. So it’s both having like, obviously, and again, a lot of what we’re trying to do is solve a lot of problems. Like for example, the caddy is obviously helping you down the golf course, like knowing obviously there’s bunkers, there’s water, there’s how far it is that the wind’s doing all these things. Which is useful, but then there’s a lot of with these language models and different technologies we’re able to implement. It’s also making it more enjoyable and fun and engaging, right? Because we can add in these personalities and these people you actually have a connection to on, cause you watch them every week. So that’s really exciting to get to do.
Topgolf.Club: Is it in your plans, I’m assuming, to bring a subscription model per month or per year like the other AI services? I mean, right now I haven’t seen it yet, but probably… It is not offered yet, but in the future, I guess the business model also going to transition into that direction, right?
Peter Bergin: Okay. Yeah, so, sure, eventually we will, yeah, for sure.
Topgolf.Club: Okay. Were there any funding rounds or private investment opportunities until now or in a Kickstarter case or something similar?
Peter Bergin: So we did a private round with the partners we have now on the cap table to kind of kickstart the business. We’re actually going into now a new round of funding. So obviously looking for investors and going through that process now with a lot of other conversations, which is exciting to be able to increase growth and obviously spend money or be much more aggressive in terms of the marketing user acquisition side of stuff.
And since we’ve been able to build a lot like the product and we really want to ramp that side up as we kind of release the new features as we’re coming in, where the caddy is like a big thing that a user and certainly got some golf courses around the world. We can use it in any language, practically any language that golfers golf in. So therefore we obviously can be much more aggressive in terms of international expansion. So looking to, so we’ve done a small round and obviously looking to do another one now actually.
Topgolf.Club: Will the second one also be in private route or will it be a public access in any level?
Peter Bergin: Well, public access meaning, like, obviously, we’re welcome to the conversations with anyone, right? So we’ve got a lot of venture capitalists and high-net-worth individuals and stuff like that.
Topgolf.Club: I meant like, oh, if the regular people who are interested in an app can invest as well for equity, like Kickstarter or something similar, but okay, nice.
Peter Bergin: Yeah, I mean, obviously you can bring individuals on the cap table as well, right? So especially early days looking for people that are aligned in terms of the interests, right? Industry professionals and stuff like that, right? Not only are we just bringing money on, but also bringing up people that are also strategically aligned.
Topgolf.Club: Makes sense. I want to talk about the customer feedback until now a little bit. Did you receive any surprising insights from customers that would say, Hey, maybe you can add this feature too, or, you know, there was something that the customers, or I want to say not the customers, but the users are, were mentioning to your support. So you like basically gathered the feedback and develop the app even further.
Peter Bergin: Yeah. It’s interesting, it’s interesting always getting feedback from the users, which is always wanted and welcomed and what we want. It’s very difficult early days, especially, because no one, and this is a positive thing, no one really knows what is possible and what we want, right? We kind of all have this idea of like, hey, this would be cool, that would be cool. But then what will sustain value? What will be able to provide, right? What will be unique? What position? So it’s always like, it’s a lot of intaking ideas, which is very helpful. And then obviously trying to filter through that, trying to understand. With this feature idea from this user, what are they actually looking for, right? They’re looking actually to improve their golf game. With this feature, is there a different way we can look at it and do it from our angle rather than just that way and maybe that way is the right way. So that, I think overall, one of the big things in life, it fits within the thesis of what I think sports are about. I think sports generally are about connections and stories, right? So if you take away one, sports is kind of useless in a sense, right? If it was just about the stories, but I couldn’t experience that with my friends, family, it would kind of be dull and I probably wouldn’t watch much of it or engage in much of it. So actually a lot of the feedback is around the social aspect of golf, right?
So can I play against my friend, like virtually, let’s call it, right? Like you go out and play around. Can I play against you somewhere on a golf course, I’m in Australia, you’re in Turkey, right? So like all the other questions. So a lot of it’s around the social side of it, which is difficult in the sense of, cause I, again, like we talk through you, we talk to users and I don’t think we, they fully have an understanding and we don’t have a fully understanding of like what that looks like exactly. Right. It’s difficult.
It’s like, is it playing against you on the golf course? Is it just seeing your scores? Is it like Strava? Is it like the social media? Is it, it’s very different aspects. So yeah, we’re trying to navigate that. I mean, but it’s again, it’s a huge aspect of golf is like sharing it with other human beings. So we’re trying to navigate how we trying to fit that in. We have some ideas, but that’s generally one piece of feedback that keeps coming back and back is, is like the social aspect of it. Which is always interesting.
Topgolf.Club: I also wanted to ask, is it possible to use the datasets with Golf AI, which the rounds you are playing with golf simulators and not the actual golf course itself?
Peter Bergin: Yeah. So another thing we’re trying to explore with obviously a lot of the simulator companies is the hope is a bit of simulators went recently. Always trying to figure out cause that’s becoming like a, that’s a, it’s a rather growing industry of golf, right? People playing, I think what it’s called, it’s a massively growing market. It’s a non-traditional golf, right? Like that’s like the short or three courses.
Topgolf.Club: Off-Course Golf?
Yeah, it’s a massive market, right? So how can we kind of fit into it? Is there a way we can fit into that? So obviously we were in discussions with some simulator companies in terms of how we can make that work, how we can make, cause obviously for them, they’re looking for more engagement in terms of their usage, right? So if we can create more engagement, that makes people want to go back. In various different ways.
Like a lot of the idea is, is can we build in the, even for example, going to the simulator, you get your swing data that would improve obviously that it brings down to the caddy of the golf course, like that improves that model a lot better anyways. So it’s trying to, yes, navigate again, that market and trying to figure out exactly what that means.
Because I mean, a lot of times. Like, I mean, the tech in the simulators are so good now, and they’re so engaging and it’s all kind of right in that house. So yeah, we’re trying to figure that out as we go, but in some discussions with some companies to try to figure that out together because. There’s no point in like, we’re not looking to build simulators anytime soon. So it’s partners, the companies that are best at that. And then we can provide obviously tech and services and users in the areas that we excel at.
Topgolf.Club: Understood. So, I mean, in the last section, I want to talk about a little bit the future plans and the vision. I want to ask like, what is next for Golf AI? Are there any exciting updates on the horizon? And also, do you see expanding Golf AI into other sports or related industries?
Peter Bergin: Yeah. Yeah. It’s a good question. So, immediately in terms of Golf AI, I mean, the caddy’s coming out soon, which is really exciting because obviously as a golf company, having somebody that you can use on the golf course is great. Rather helpful. So that we’re really excited about. That’ll be like essentially one of the biggest features of the platform to be able to use caddy on the golf course. Really exciting and it’s really been using it and super helpful, super fun to be able to use. And then from that, I mean, there’s various different opportunities. We’re actually in, which we have a club or something we haven’t touched on is with obviously AI.
There’s a lot of opportunities in innovating on the user experience of various different touch points in golf or anything in sports in life. So even like on the tee time booking side of things, just being able to build in these AI agents, let’s call them. Rather broad term that people throw out a lot, but being able to call a phone number and book a tee time anywhere in the world, like that would be like, obviously, a path we’re moving towards. And talk to a lot of people to try to enable that. We have some demos that we’re actually demoing right now with people and testing.
So being able to tap into the tee time market is really exciting. And I mean, in terms of, which we haven’t touched on a ton, we’re coming out with some new features on the media side of things and innovating on the user interactions of the media. Obviously it’s a rather broad sentence to say. So it’ll be much more easier to visualize once we come out with it. That translates very quickly and scales very quickly to every other sport in the world. So there’s obviously discussions as we do as entrepreneurs and startup founders, we tend to have the mindset we can do anything.
So trying to navigate the excitement of being able to do it across all the sports. And this was the discussion as we go back to the inception of Golf AI. We had this discussion of, do we do many different sports and one very specific sport? Because we saw the opportunities in both and which one do we want to do?.
And that’s why we chose golf specifically because we can do it in the verticals of both while you play and while you watch. So the media stuff while you watch, while as a fan, that scales very quickly across many different sports and obviously it has implications in betting as well. Because obviously like media information is essentially important in betting. So we’re innovating a lot on the user experience side of things. And like, for example, I’ll virtually touch on it.
Then the hypothesis is, or the general gist of the information is, there’s the way we consume, let’s say stories now is generally through headlines, and headlines, there’ll be a ten headline that headlines give you a little bit of a sneak peek in terms of like what that story is about. And then you click on what’s interesting. And then you read that article, maybe go back and read another one. And even like in podcast, you’ll get a list of the podcast, then you’ll click on and listen to what I see it as is that is like a window into the story and like that’s showing you a little tease of what’s interesting into the story. But generally in sports, our first touch point isn’t the articles, isn’t the headlines.
It’s actually like the scoreboards, the scorecard in golf, the leaderboard, that like kind of very quick, beautiful view of the stats, what happens. So what we’re doing is putting the story right at that touch point. We’re going to show you what’s interesting within that. And then if you click that, it’ll show you the story of what happened. Like, for example, you can click on the fourth round of a player that got 10th in a tournament. Maybe he was ranked 500 and got 10th. So that’s interesting, right? All of a sudden you click it, get the story and audio form. In text form or anything. Going down the route of even like things like interactive podcasts, right?
Because we can do this with the technology that we’ve built on the media side. As a podcast starts, you can then guide the podcast into various different ways and allow it to navigate and move and shift into different topics and different voices and call into the podcast. So there’s a lot of space within that to go across the different sports. Cause I didn’t imagine kind of like falls into like the while you watch category. And we can scale it across different sports very quickly, but we’re obviously focused very, very intensely on golf. And we’ll see where that takes us I think for now, but that’s kind of the big plan.
Topgolf.Club: Yeah, sounds really exciting and I will be, and I think also my audience will be following you guys and your story this year and upcoming years. Thank you very much for joining us today, Peter, and thank you for your time. And it is really nice to meet you personally.
Peter Bergin: Yeah, I appreciate it. Awesome. Have a nice day.
