3517: How AWS and the PGA Tour Are Changing Live Sports Technology

[00:00:04] Have you ever stopped to think about what it takes to track every moment of a golf tournament? One that spans 200 acres with more than 30,000 shots in play. This is the kind of thing that I think about and it's a question that I was pondering today as I sat down with two people who have a front row seat to one of the most ambitious data and AI partnerships in global sport.

[00:00:30] So in today's episode, I'm going to give you a chance to hear how the PGA Tour and AWS are together reshaping how fans understand the game, follow their favorite players and even experience the sport in real time. In a moment, I'm going to introduce you to two guests and together they've been working side by side for years, building systems that turn raw shot data, historical archives and generative AI into personalized multi-language experiences

[00:00:59] for millions of fans. And their work stretches from automated play-by-play to automated camera switching to a mobile experience that adapts on the fly to what every single fan wants to see. So today you can expect to hear how they think about accuracy, scale and validation and storytelling, along with a much broader shift toward AI agents and hyper-personalization.

[00:01:27] And if you follow any live sport, this conversation should give you a glimpse of how fast things are working behind the scenes. So what happens when a century of golf history meets a cloud platform that wants to reinvent how fans watch the game? Before we go into today's episode, I just want to give a quick shout out to my good friends at Denodo. The data world is louder than ever.

[00:01:56] Yeah, AI hype, lake house complexity and pressure to deliver more with less. But my friends at Denodo, they're helping enterprises make sense of it all. Because their logical data management platform provides a unified data foundation for trustworthy AI, lake house optimization and data products that finally bring self-service to life.

[00:02:18] So combined with AWS, teams can now access secure, governed and AI-ready data that accelerates every step of the journey. So whether you are a CIO or a builder, Denodo and AWS can help you unlock real AI outcomes without the headaches of data replication. And you can learn more at denodo.com slash AWS. But now it's time for me to officially introduce you to today's guest.

[00:02:49] Welcome to the show. Can you tell everyone listening a little about who you are and what you do? Yes, I'm Eric Hanson. I'm the VP of product at the PGA Tour. So all of our fan-facing platforms, the PGA Tour.com, our mobile apps, our fan-facing AI initiatives fall underneath me and my team at the tour. And we're very fortunate we've got not one but two guests today. So could you tell everyone listening a little about you two? Yes, I'd love to. My name is Elaine Chasen and I lead the global golf team at AWS.

[00:03:16] So I work with the premier golf organizations around the world and I've had the privilege of working with PGA Tour for the last eight and a half years. Fantastic. Now you've both been in Vegas during one of the busiest weeks in the tech world. So before we get into the heavy stuff, how are you finding the mix of late night keynotes, AI announcements, and the overall pace and crowds of reInvent? I think it matches the changes you see in the world right now.

[00:03:43] I think if you see how fast we're moving in the AI space and you see how quickly AWS is leading in that space, I think the energy and the atmosphere here matches that. And so if you go to a lot of tech shows, you don't see people moving as quickly all the time, especially for early mornings. I get up early and I see people just so excited. We're out late at night at dinner. You see people still going really strong there. And I think it just matches the excitement that people have with the technology that's coming out right now. Yeah. And are you saying very similar as well?

[00:04:12] I'm in the reInvent bubble for many days. I don't even know what day it is, what time. I feel like we don't sleep here. But yeah, I mean, plus one to what Eric said. I mean, it's just a great energy. Like you just don't feel tired because you're energized. We've had a lot of amazing conversations, not only with AWS product leads, but partners and customers, a lot of customer collisions. And it's just been fantastic to hear the conversations. Excellent.

[00:04:38] And one of the things I try and do on this podcast every day is get people thinking about how technology impacts areas that you don't automatically associate with technology. For example, we're talking about Formula One. And for the most part, I had no idea. But at the end of the podcast, I learned that there's 1.1 million data points on each car every second, which blew me away. So tell me about PGA Tour and why technology is so important there too. It's very similar to what you said with F1.

[00:05:09] If you think about our sport, we have 18 holes that are played over a course of 150, 200 acres. So that's basically 18 different arenas that we have running concurrently during the course of an event. Not only that, but you have sometimes 144 or 156 players all playing in the event that week as well. And so you have all these different fields of play. You have all these different players. And then their position affects another person on the leaderboard as well.

[00:05:36] So much like you said with F1, when someone goes into the lead, it affects everyone else on the leaderboard. We have the same thing. So if someone makes a birdie on a hole, they move up the leaderboard, it pushes everyone else down. That leaderboard. So you have to track all that information. Not only that, we have to track every ball that's hit across the course. And that sometimes is over 30,000 different shots during the course of a PGA Tour tournament. So you talk about F1 and the amount of data points there.

[00:06:01] We have just 30,000 just latitude, longitude points that we have to luck at over the course of an event. Never mind the fact that we're tracking every single ball off the tee using radar. So we're tracking all these different shots and trajectories that go down the course. We're tracking different statistical components of those shots as well. And so they add up over time, over the course of the event, which leads us to having a lot of different data points stored inside our systems.

[00:06:28] And when you think about the PGA Tour's long partnership with AWS, how would you describe the shift from maybe traditional broadcast workflows to a world where 30 years of media guides are now searchable data that is powering fan experiences? Another example, I was talking with Real Madrid and Man United, and they were talking from a soccer point of view. They've got 60,000 people, 70,000 people on the stadium there that day, but they've got 5 million viewers around the world that they need to improve that fan experience.

[00:06:57] Presumably exactly the same for yourselves. So absolutely. And, you know, with PGA Tour, you think about just the massive amount of data that we have. So we have 100 years of video assets, their media asset management system that they've migrated into AWS. They take all of the shot data. So they have a proprietary system called ShotLink, which captures all the data that Eric was talking about, every single shot. All of that is in AWS.

[00:07:25] They shut down all their data centers and their colo. All that runs in AWS. So you have this massive media lake that they have that then they can leverage that and apply AI and generative AI to be able to quickly give the fans what information they want when they want it. It's incredibly cool. And as you mentioned a few moments ago, the tour deals with 32,000 shots across 200 acres every day. So how does that scale push you towards automated commentary?

[00:07:54] And what gaps did you want AI to help you close as well? Yeah, I'll start with the gaps part first. So when you have that many players, it's impossible to write content for every single one of those players. It's also really important to put context around every single shot that happens. So during the course of an event, we only have about 30% of all those shots on video. So 70% of those shots have no video to actually show what happened. So that's where generative AI comes in.

[00:08:22] We can use Gen AI to basically provide context around that shot. 300 yards. We can tell you how frequently are they hitting the fairway this week? How are they comparing to the rest of the fields? There's no commentator there to say that, but we're adding in that commentary inside of it. So that's where Gen AI helps us with that part. Also, I mentioned we have 144 players in the field. We need to be able to write a recap of the round so that if you miss what happened, you can go back and read and see what happened.

[00:08:51] And so we don't have enough writers to write stories every single day about that. And so by a push of a button, a player comes off the course and we were able to write a story about how they did on the course that day. Well, also, I think what the PGA Tour does so well is they really take input from their fans. So they have this initiative called the Fan Forward. It's over, what, 50,000 fans that they ask questions. What do you want to see? How do you want to see it? Like, what information do you want?

[00:09:18] And then we use mechanisms and programs within AWS to help them solve some really hard problems using our technology. I mean, one example is the AWS Generative AI Innovation Center. So we come together with the tour, we ideate, we talk about the problems, and then we work backwards and then use our technology to help them solve it.

[00:09:39] And also, I was reading before you came here today, the automated play-by-play system and how it uses a fact prompt and a context prompt to create commentary that ultimately feels more natural. So how did your team balance speed, accuracy, and storytelling quality that the fans expect? Because if anything's slightly off, they will pick you up on it, won't they, straight away? No, you're exactly right. And so when we were coming up with this concept, the original version of it only really talked about the fact part.

[00:10:07] And that's essentially, if you look at a lot of Gen AI work that's in sports in particular, that's what you see is it'll tell you a fact. And we said, that's not good enough because we need to actually put context around it. And so what we end up doing is we take a fact and then we have a rules engine that basically says, based on this fact, go look at these other different statistical points and add context to it on top of that. And so that way we're enriching the data so that fans have more information to know.

[00:10:34] And the way we use that, the way I use this personally, is on Sunday, we're down to two or three players on the course. Say there's a one-shot lead. A player has an eight-foot putt. We put on there, on the AI commentary, that so-and-so has an eight-foot putt for birdie. He generally makes an eight-foot putt 32% of the time, which is better than or worse than the PG tour average.

[00:10:59] So I look at that going, okay, well then, you know, he's probably got a two out of three chance of missing this. And so we're likely not going to see a lead change. And if he makes it, you're like, oh, wow, that's a one out of three that he just made right there with it. So to me, it adds more context to just show how good these players are because making an eight-foot putt is really hard. And seeing that guy actually sink it, that just shows you just how good they are.

[00:11:23] And for many fans listening, they're going to be curious about the next step, which I often read about is audio commentary across every shot live. That's something that cropped up in my research. But how do you view the opportunity to do things like add multiple languages or real-time audio for every player during a tournament? It feels like there's a lot of opportunities there. There's so much we're doing together. It's like we – I think that is really one of the things that works the best for our partnership. We push each other very hard, both of us.

[00:11:52] I mean, they drive our AI roadmap. I can't tell you how many feature requests they have. So they push us very hard. But we had – I'll give you an example. This past year at the Bay Current Classic, which is in Tokyo, just outside of Tokyo. So we took what PGA Tour had, their AI text commentary that Eric mentioned that gives statistical information and color commentary on every single shot. That was in English.

[00:12:16] And we just – a couple weeks before we went to Bay Current Classic, we were executing a hospitality program there. And our AWS marketing team said, can we have instantaneous language translation? And can you put it in production? And so the PGA Tour stepped up. You know, the team, they've built an incredible software development lifecycle, how they actually do development.

[00:12:41] And they delivered in production instantaneous language translation in Japanese as an example. And so one of the ways we're expanding upon that then is trying to take it to the next level and actually do automated TV production. And so this week at the reInvent Pro-Am here in town, we actually practiced actually automating a live sports broadcast through AI automation. And the way that worked was we have a camera behind our tee.

[00:13:10] So when a player hits their first shot, within a second, we know where the ball is going to land off the tee. And so we have three to four cameras around the green using the AI technology that AWS helped us create. We are able then to choose which camera is going to show the ball land the best all through automated switching. So as the ball is in the air, the camera switches automatically to the one camera that's going to show it where it's going to land the best.

[00:13:35] And then it actually pans and tilts and zooms on the ball automatically as well to show you where the ball lands. Then we actually had a leaderboard that showed closest to the pin and all the graphics were automated as well. So this entire production was running with no humans involved, cutting between cameras and putting all the statistical graphics on the screen, again, with no people involved inside of it.

[00:13:57] You know, and to give additional context to that, which is so important for the fan, why this is so critical for the PGA Tour, it's like one of the most exciting things that we've had the privilege of working with them on is, if you watch a regular broadcast, you know, watch a golf tournament on TV, only about 30% of the shots are actually televised. So if your favorite player is not in the leader group, you might not get to see any of their shots.

[00:14:21] Well, using this technology, we'll be able, the PGA Tour will be able to capture every single shot and maybe have a second screen experience or however they're going to present it. But they'll be able to capture all the shots of all the players at a tournament. Wow, that's incredible. And of course, TourCast already lets fans see shot shape, radar data and progress around the course. How do these AI layers change the way fans can follow their favorite players, especially during long days with simultaneous action across the field?

[00:14:50] I mean, you made a great point about being able to follow the players individually, but anything else? So we're going to be rolling out in January a new feature on our mobile app that solves that exact problem that you're talking about. So we have a section on our mobile app called Your Tour. So your favorite players now are going to have a lot of this Gen.AI content surface to the very front of the app that depending on where they are at the time of the day, it's going to serve you up different Gen.AI components. So it's either going to give you that round recap.

[00:15:18] It's going to tell you about the hole they're on and the shot they're taking right now. It's going to serve up different video clips for you as well if they finish the round or they've hit some really good shots during the round. So we're really leveraging all these Gen.AI capabilities and all the things that we've built here to really personalize the fan experience in our mobile app for people based on their favorite players. Now, we're mentioning generative AI here, and it's not only changing commentary, it's also changing fan creativity.

[00:15:43] And the course creator at the Tour Championship lets fans design the entire golf course in just a few seconds. So what did that experiment teach you about how fans want to interact with the sport? Did you learn anything from that? Yeah. I mean, I think what we hear, at least from fans, is you have different types of fans, right? You have some like my dad who likes to watch the traditional broadcast. Don't mess with that. He likes it the way it is. You know, Sunday, it's always on at the house.

[00:16:12] But then you also have, you know, younger fans that want to experience golf in different ways. And I think what's incredible about PGA Tour is we are working with them to experiment, you know, different ways that create a classic, you know, different ways they're going to enjoy content. And, you know, the thing is about the tour is they don't have staff to do, you know, just massive amounts of work. They're a relatively small team.

[00:16:36] And so I think if you use technology and automation and use generative AI to help them test and experiment and find different ways to reach the fan, they just like the example I use with Bay Current Classic and instantaneous language translation. They've built their teams. Their tech teams are at such a high level that they can experiment and try things and sometimes trying things in production, which is a little scary. But and then they they do a great job collecting metrics and seeing what the fans like.

[00:17:07] Now, as an ex-IT guy, I've got to ask, we often talk about the risks when AI is integrated, especially into live content. So is there anything you can share around the kind of validation standards that you set? And how do you ensure that the commentary stays aligned with the PGA Tour accuracy requirements? I know it's a bit of a gray question. It's an important one, isn't it? No, it's a great question. It's something that we bake into everything that we do. So a core to everything that we do is validation.

[00:17:34] And so we've even created validation agents to be able to make sure that when we build different capabilities, that we always have validation at the core of it. So we talked a lot about our AI shock commentary. And you're right. There's a lot of things that could go wrong with that. And so what we built into that is validation. And essentially, we built a validator that looks at the commentary, make sure it matches all these different components, brand voice, right? Statistical information. If it passes, it goes on.

[00:18:01] If it fails, we just send it back to the LLM again and have it reprocessed. Because usually, if you send it back through again, it's OK the second time. And we essentially have four or five different grades that we give the commentary. And that we essentially, if it continues to fail, we reduce down the commentary part of it and just deliver the fact part of it. And so that way, we're actually protecting ourselves from having some of these hallucinations or some of these incorrect pieces of information go out.

[00:18:28] Because we're basically self-metering all these different pieces of content on the back end to make sure we're delivering good, valid content to our fans. I think that the tour has done an outstanding job with LLM Ops. In fact, we published, co-wrote a blog together, PGA Tour and AWS, talking about the CloudWatch dashboard that they built on Amazon. And how it showcased even with the Players' Championship back in March when AI text commentary was launched.

[00:18:57] They had this rules engine that Eric talked about that really was, you know, automatically checking and making sure that what was published was accurate. Because it's everything for their brand, right? You can't have misinformation out there. But they've built this automation rules engine on CloudWatch that is, I think, a lot of our customers could learn from. Because several customers are still kind of at the early days of generative AI. They're doing proof of concepts. They're experimenting. They're trying it.

[00:19:25] But they've actually built an ops model into what they deploy that is, you know, protecting their brand. And with predictive elements that you've both mentioned today from where the ball will land being generated in just 1.5 seconds. I'm curious, from both of your perspectives here, how do you think these insights will possibly change how casual fans and broadcasters, how they talk about the game and analyze the game?

[00:19:50] Yeah, I mean, I think, you know, as the PGA Tour launched the World Feed back in March of this year. So they are looking to, in the future, customize the World Feed and maybe have localized graphics. You know, really, if there is, if we're in certain markets and then these markets have PGA Tour players that are from those countries, you want to really personalize this. You want to hone into these, you know, players.

[00:20:17] And so I think what PGA Tour is building for the World Feed is going to be really exciting over the past couple of years. And we're happy to be a part of it. Just to add another example on that. In March, we met with our Japanese broadcasters. And we were showing them the text commentary system that we had created. And they asked us if they could use that on TV. And they said, we're following a lot of Japanese players.

[00:20:45] We don't necessarily have all the stats and data about them because they're not necessarily part of one of the lead groups each and every week. But they said, can we use this data to enrich our broadcast? And the commentator told me, he goes, this will make me sound smart on TV because you're surfacing a lot of content that I didn't know. And he goes, I could go look up all this different data. But he goes, this comes up so much more quickly than me going through and looking at the stats.

[00:21:11] He goes, I can just read off some of these things to the fans through TV to really enrich their experience. Incredibly cool. And finally, as we look ahead into 2026 and beyond, where do you both see the biggest leap forward as you continue working with AWS? Whether it's fan personalization, multilingual experiences, virtual assistants, or something we've not talked about today. What would it be? I think it's all of them.

[00:21:37] I think the way we look at it is we progressively go into each of these areas. You highlighted a lot of pieces, the Gen AI part, the personalization part. We're going to continue to drive into that and get good at that, which allows us time then to go focus on the next big thing that's coming up. For us, it's the agentic side of it. It's like, how do we create more of these agents and then orchestrate those agents together to build more of these different capabilities we need? We're also looking at how we improve employee productivity with this.

[00:22:05] There's a lot of manual work that happens within our building, within our office, that these things that we're building right now for fans could also help our internal employees with it. And so can we start going into the productivity side of the business as well and saying these agents can actually help you with other things to make your job easier and better? I think for me, one of the things I'm excited about just being a fan of golf is what the PGA Tour is doing around hyper personalization.

[00:22:33] So Eric touched on it briefly before, but your tour is something that's going to be coming out next year that I can actually on my phone personalize and have content that's interesting to me. And not just about the players I like, but what statistics I like. They have an incredible amount of data and they publish all that on their website and their app. But the only difficulty sometimes right now is it's just so much to consume.

[00:23:02] I mean, they're playing week in and week out, 40 tournaments, like you mentioned, 144, sometimes 156 players. It's a lot, sometimes overwhelming. So I'm excited about that hyper personalization that they're going to be building on AWS and their app. So stay tuned for your tour. Oh, awesome. Well, I'll add links to both of your LinkedIn's. But if there's anybody wanting to just keep up to speed with some of those tech developments, is there anywhere in particular you'd like to point everyone listening where they can keep up to speed?

[00:23:31] I'll point you to the blogs that AWS writes because they do a fantastic job doing deep dives into this, especially people who are really technology focused. The AWS team does a fantastic job documenting a lot of this. We have a section on our website on pjatour.com that talks about what's new. So that kind of gives you the fan-facing side of it. But the really deep down tech parts of it, AWS does a great job of documenting that. We have a PGA Tour landing page. So if you go to – I don't have the URL off the top of my head, but just Google it.

[00:24:01] It's the AWS landing page for PGA Tour, and it has case studies out there. It also has videos of some of their speaking engagements that they've had. So we talk about a lot of this capability. A lot of the blogs that Eric mentioned are on our landing page at the AWS website. Awesome. And to save everyone at Google, I'll add links to everything there as well. But more than anything, just thank you for stopping by today. Really appreciate your time. Thank you. Thank you very much.

[00:24:29] I think conversations like this remind me of how quickly sports technology is moving and how much of that change is shaped by collaboration, collaboration that sits well out of sight. And I think Eric and Elaine shared a clear window today into the foundations behind it. Whether it be validation layers that protect the brand to experiments that bring every player into view,

[00:24:55] rather than just a few that are selected on a broadcast feed. And it's a shift that speaks to both the ambition of the PGA Tour and the pace of innovation at AWS. And I really think that there's a wider story here about how fans will consume sport over the next few years. And hyper-personalized feeds, multilingual commentary and agent power production pipelines. All these things are edging closer to the mainstream.

[00:25:23] And it's this tech partnership that is already proving what is possible. And I walked away from our conversation today almost sensing how many ideas are still on the horizon and how quickly they will move into production once those foundations settle. So if you follow golf, cloud engineering or just interested in the future of AI and live entertainment, I'd love to hear from you and what stood out for you in this conversation today.

[00:25:52] So where do you see the next wave of innovation arriving? And how would you like these experiences to evolve? As always, techblogwriter at Outlook.com, LinkedIn, AX, Instagram, just at Neil C. Hughes. Loved to hear from me on this. But I'm afraid we're out of time once again. So I'll be back again tomorrow with another guest. But thank you for listening as always. And I'll speak with you all again nice and early tomorrow morning.