Scroll Stoppers: what's real anymore? The AI revolution in sports content

0:03Hello and welcome to Scroll Stoppers, a podcast series from us here at Leaders in partnership with Tentos. My name's Henry Breckenridge. I'm the Senior Content Manager here at Leaders. And I'm Ben Weisfeld, Founder and CEO of Tentos. In this episode of Scroll Stoppers, we'll be exploring how AI is reshaping the way sportsite organizations create, distribute and scale content. We'll talk about the tools and the ideas that are pushing the boundaries and rethinking what's possible across sport. And to do that we're joined by Scott Gutterman, the former SVP of Digital and Broadcast Technologies at the PGA Tour and now currently Chief Digital Officer at NextLeague, the AI operating partner for the sports industry, where we discuss the opportunities, challenges and realities of AI generated content in sports.

0:51Up next is Scott Gutterman. Okay, as trailed down the line is Mr. Scott Gutterman. Scott, thanks for joining the Leaders Worth Knowing podcast. How are you doing?

1:02Scott Gutterman:Good, good. Thanks for inviting me. No worries at all. Excited to get into this one, chaps. Time to strap on the AI boots and given your expertise in that area, both Ben and Scott, excited to delve into it. So obviously AI, a universal theme kind of inside and outside of sport, you know, you can't move without hearing about it. And today, keen to get into AI content generation specifically. But the first question, which I haven't teed you up for, is I'm keen to know what your last prompt was for AI, either personally or professionally. So Scott, I'm going to put you on the spot first. What was the last prompts that you put into AI?

1:50Scott Gutterman:I'm laughing because I was watching the, here in the States, I was watching the Florida State SMU game last night on ESPN. And I'm a Florida State graduate. And so I was actually forming up a chatbot that would give me game updates in the words of Coach Bobby Bowden, who was the legendary coach at Florida State, and Gene Decker off, who was the play-by-play guy, and seeing what kind of feedback I would get during the game because I was so frustrated with the outcome of Florida State losing. And were you satisfied with what you got back? It was. This is the thing that amazes me, and I'm trying this when I was on the tour and other times.

2:32Scott Gutterman:The data and the commentary I was getting back was very topical and very live, which is very different than six months ago, which is really interesting. And that was just using a general cloud subscription. That was it. Very good. And Ben, same to you. What did you last insert into AI? On the train on the way here, we're in the process of building out a sort of social media playbook. And I've been trying to figure out how to get the phones aligned with the feed that I needed to pick up. So I was playing around with how I could try. And I do a lot of stuff on Claude Code all the time. So I'm fixing bugs before we deliver it to a client next week.

3:13So it was a bit of a building one as opposed to finding anything out or anything like that. I do completely agree with what Scott just said, though, in terms of the reaction, the liveness now, particularly on all of them, is just gone through the roof. Like you can get stats like straight away. Indeed. Which is great. Scott, you made the move to Next League earlier this year. and I think you're probably coming up to your six-month anniversary or maybe it's just happened.

3:42Scott Gutterman:I'm a little longer than that now. I'm starting in January, so I'm about eight months now, I think. Yeah, about eight months here. Congratulations. Thank you. You were previously at the PGA Tour for two decades, I believe. Tell us about you were kind of in the digital team at the PGA Tour and you kind of made your way up, I think, to SVP of broadcast operations, I think. Yeah, broadcast and digital operations. Yeah. Could you tell us when AI started to sort of take root at the PGA Tour, you know, when you reflect back on your time there, what kind of moment did you sort of feel that AI really was becoming apparent within the fabric of the organization?

4:23Sure.

4:23Scott Gutterman:So look, for us, and I think a lot of people in sports, AI is a very broad definition. So in the sense of AI, you know, machine learning, like it was 10 years ago with our shot link systems and things that we were doing internally with all of our systems, particularly around scoring, was a very big part of what we were doing. So a lot of algorithms were developing win probability algorithms, we were developing new stats. So using AI to develop statistical information was a very foundational thing that we did at the PGA Tour. I think the question that you're asking is when did generative AI take root at the PGA Tour?

5:00Scott Gutterman:And that really was about, you know, what's this? This is 26. So about three years ago, about 2023, when things started to emerge and the models started to emerge in a more commercial manner, our mission at the PGA Tour was to tell a story about every golfer. And so the question was, with, you know, 100, 150 people in the field, how do you do that? And we were really, when I ran the content teams, the struggle was we could always report on the top five to ten people and a few other people occasionally, but you weren't going to hire enough people to cover all of our players. And so what we really began to see in the early days is like this could be a tool for us to tell a story around every single player that's on a course and in a wide variety of ways to do that as well too.

5:46Scott Gutterman:And we began to see that in work we were doing with WSC Sports, we were generating video clips that were going out to all of our fans and then very closely working with AWS as they began to develop bedrock and begin to work with the very early versions of the model. So we very quickly moved to a lot of prototypes in 23 and the beginning of 24. But in 24 and 25, I said, look, everything we do has got to be operational. And let's start with telling a story about every single shot on the PGA Tour. And how do we do that by saying more than just this player hit a shot, you know, 300 yards, they've got 125 yards of the hole.

6:23Scott Gutterman:What I wanted people to know, what the team wanted people to know is what did that shot mean based on the rest of the event and that was really the beginning of us developing a wide variety of agentic systems um some very early agentic systems that would help us create content and in many many different ways yeah so it kind of sounds like you had the sort of uh democratization of storytelling with i think it was tourcast um where to your point you know you were able to sort of explain the narrative and the context of each shot and why it mattered not just Rory McIlroy's got a 10-foot putt to win the hole you know what's the likelihood of McIlroy making this putt um give me some historical information and then secondarily uh it sounds like it was a um you know an access point as well with every shot live you know 150 players on the field 200 acres of golf um sorry 200 acres of land uh across four days of competition, giving fans an opportunity to watch that stuff and consume it as well.

7:25I'm intrigued to know, Scott, how your perspective is now in the industry around AI content generation. Obviously, you were at the PGA Tour and you were obviously immersed in whatever you were doing there. But now you're on the other side of the fence, so to speak, at NextLeak. you know you're servicing clients like the pga tour and other rights holders um ben going to come to you in a minute on this one as well like how how has your perspective perhaps changed if at all in the last sort of eight months as now you're kind of i guess taking a bit more of a bird's eye

8:02Scott Gutterman:view of the sports industry sure so i think for us at next league it really is about what can we do for fans of our clients, so whether it's NASCAR, the PWHL, or USGA and others, is how do we help them tell their stories? And what do we do to enable them to focus on the things that are unique about their business? So in a little bit, it's similar to what we did the tour, which is like, what are the things that we need to do every day? And can we automate those out in a way from what the teams have to do. Teams shouldn't have to write stories about how to watch. In some cases, like a basic gamer, you really don't need somebody to write a basic gamer.

8:45Scott Gutterman:The summary after every round, particularly for golf or NASCAR, it's how do we talk about how many laps this driver has made and where they stand in the race. We want to be able to use the technology to generate that information so that fans have that. And in a wide variety of ways across a wide variety of platforms. But really what I believe it does is it loosens up the content teams to go tell unique stories, to go find something that is not something that you're going to see just based on the action happening in a sport. So tell the story about a family that may be watching a driver for the first time on the track and how they're performing.

9:25Scott Gutterman:You want to focus on the family. You want to focus on what the driver's talking about because that is what I think the main interest is going to be in growing the sport and growing the attraction of the fandom for that particular athlete or that driver. Yeah, absolutely. And Ben, just coming to you, what's your kind of take at the moment in terms of how you are seeing AI content generation with rights holders? Are you seeing much kind of differentiation with your clients across the board or yeah, just give us your kind of overall assessment of where we are. It's really interesting. I think we went through a phase probably six months ago where everyone was just like oh my god what are we doing with ai and and everyone wanted something but they didn't know what they necessarily wanted i think now where we've already shaken out across our board is some clients are very very risk averse and they they don't want the content to feel fake or inauthentic and all the buzzwords that everyone's had for years and years uh all feeling like it's real or whatever you've then got the risk averse then you've got the complete alternate end which is where those ones are having fun with it testing it it's it's less restricting to to go and create i don't know uh 30 player profiles and turn them into animated characters and have them do whatever you want and create loads of content like they're just there's there's very little risk for certain rights holders versus others where they're terrified about ip infringement and the broadcasters and what so so i'm seeing a really interesting mix the ones i think are probably getting it right somewhere in the middle where you're using it from a content creation perspective to help prove a concept or develop an idea or stress test something before the players turn up the ones that aren't there i think in some way shape form they are getting left behind whether that be they're still spending loads more money on actually having to do everything for real or they're not taking the leap and trusting an agency to go on and just try it and make it or they're taking weeks to sign contracts because there's an ai clause and then they don't quite get it like there's a real mixture of everything i think to be honest it's just the same as the man or woman on the street like some people are really into it they're using it in every single part of their daily life their emails are being served to them in a summary first thing in the morning on claude then they're doing a code thing and like it is completely the same so i think it's not just rights holders some talent that we work with are really pushing us to see how we can do some really smart stuff with their ip and their likeness and what you can do there some brands are really sort of saving tons of money by going down the production route in that way so across our business we're seeing really different reactions and i can't give you an answer who's better who's right who's wrong it's just different i think It's just a different take and sort of risk averseness, basically.

12:23And you at Tentoes have been busy over the spring and summer sort of developing your kind of own AI. Ten Beats, I believe, is what it's called, Ben. So could you take us inside, you know, what that is and why you brought it to the sports industry? What problems is it solving? So figuring, talking all the way about it. So about eight, nine months ago, we were set a challenge and I was like, okay, this is the perfect use case. But going back 18 months, we did an AI week at Tentos. We had these guys come over and teach everyone in the agency the foundational sort of knowledge of all this thing. And at the time, I've always been very techie and very interested in what we can learn, how we can be better, how we can be more efficient.

13:13And we had this WhatsApp group of about 200 fans from all over the world and all different sports, all different types of nationalities, etc. And it was an absolute nightmare because basically the reason we had the WhatsApp group was if we had a question, we could just ask a global fan base and get a bit of a gut feel as to whether or not it was feeling like we're in the right zone or not. anyway when we were doing this ai week there was a whole discussion and conversation around personas and i remember like even back at uni way like you had to do like the pen profile and every brand used that pen profile was like the perfect person for who was going to be their target audience anyway from that i started to build personas and i wanted to understand obviously using gpt we did a lot of work with perplexity we did a lot of work with claudine trying to figure out how you could mimic the reaction or feelings or a point of view more importantly than anything from a persona if you gave it the right ingredients so i mean the first test was i built a thousand personas and they were all tottenham hotspur fans this is like the first i'm a spurs fan um and i started asking the personas in this little site that I'd made originally on Replit like how did that moment in a season make you feel and I had a really good idea of how it made me feel so it was actually like the perfect gut reaction to say like I was like okay and they were giving me like information that was relevant as of that morning and it was really interesting it was sort of like the defeat on Saturday absolutely killed my morale but if I then asked it on the following Monday it was given me like fresh optimism so i could understand that if you had the right creation behind the personas there's something really powerful for it and we always try and understand or second guess in many respects what fans want to see what they want to think what they want to feel when they're watching content documentaries seeing an ad seeing a post from their favorite athlete whatever that looks like so we set upon building thousands and thousands and thousands of personas um took us a long time to get to the first hundred thousand but within that first hundred thousand data sample we had really started to formulate exactly what these personas were going to be where they were going to be from what sport what country what made them what data sources we were pulling all of those things um and then we just went from there and over the last sort of six to eight months we built another 900 000 we're now beyond a million personas and we cover cover every single sport that you can think of uh the sports are weighted based on global popularity so soccer's number soccer football is number one scott's rubbing off me nfl it's got it's got it's got every sport all the way down through to judo at one percent of the fans and and we basically have built a platform that enables us as an agency and also clients of 10b to go on and ask their fans what they're thinking what they're feeling do you like this idea do you not like this idea this moment happen and it updates every two minutes so if you're asking a question 10 minutes before or 10 minutes after and something big has happened in the world those personas are going to give you a very different result so yeah that's that's the long version but hopefully gives you an idea of what we want it to understand what fans are thinking so could you give me an example then say if i was to go into 10v and i uh wanted some direction on a problem that i'm having say i'm a brand or a rights holder and i asked 10 beat you know give me uh the solution or give me um some insight on this area could could you give me an example of what i might ask that 10 beat would be able to provide which otherwise would have taken me maybe four six weeks to figure out and that was it it was the four six weeks right you would have to do focus groups and market research and all that sort of stuff so perfect example massive rights holder of the premier league is a brief that we've had wanted to really understand what the future rights packages that they wanted to buy would look like so in order to start that they needed to understand how fans were thinking and feeling across a season and what behaviors they were showing towards consumption of tech matches social all that so we built a custom persona set of about 20 000 fans global premier league fans and the question was across a premier league season how do you think how do you feel what do you do and you can get really granular with the prompt or you can get quite light with it depending on how much you can ask follow-up questions but in theory you would ask a question all of those personas then give their point of view the tool then really elegantly then gives you the exact summary the key points you then get quotes from every single persona and it's telling you exactly what they're thinking and feeling from there we can then pull out trends patterns and then obviously a human then looks over it to check that it's not completely wild but what we've done and i think this is why we've had success with it is we've tested it against real life market research asking the same questions and the results were basically identical interesting because one of my sort of challenges to uh i guess what you've um mentioned there was you know surely some people are coming to you and saying it's better just to ask humans right But what you're saying there is what's being kind of, for want of a better phrase, being spat out by 10B is actually an accurate summary of what real people are feeling anyway.

18:57Listen, I'm sure Scott sees it as well. Some clients are never going to buy into this type of tool. They want real people with real point of views and they do not care if AI can do it, that they're going to do it in their way. We've had that. We've pitched this to loads of people. some people look at me like i am absolutely crazy um so yeah it's it's not for everyone i can i can sit here and say our data is is the best data that you can put into these personas which gives us the best chance of getting the best results and the most accurate results out so yeah i think you're always going to have skeptics what ben's pointed out here too is

19:36Scott Gutterman:like you listen to this like you can listen to the investment that ben and the team have put into this, and this is a lot of what we're working with clients on too, is they have an idea, but to bring that idea to fruition, there's an investment that needs to be made. There's data that needs to be accessed. Somebody has to sit and think about how do I build one persona and then take that to 900 ,000 personas. And then on top of that, somebody has to think about how am I managing this now, right? Because Ben, you probably have got this built against some model or maybe you've got model agnostic and so you're able to switch models in and out.

20:08Scott Gutterman:And so that behavior changes. But the other side of what Ben's done, again, is similar to what we're helping clients with, which is how does that compare to the real world? Because it's like you don't – I think clients that are really sophisticated are going to want both for a little while because they're going to want to see how they both compare. So I think it's okay to do that for some time and maybe for some places that will always be the practice. But at the very least, you can kind of get some insight on the study before you actually go to launch it with real folks to really see what's going to happen and understand those behaviors.

20:42Scott Gutterman:but it took like i mean men like it took what almost a year is what i'm hearing yeah to kind of get to this system and then there'll be times that'll need to be maintained and updated and this is i think often what people people think they can open claude and uh open ai or or whatever they're going to open and just start getting these things out you know getting this type of information out of them and that is not at all what it takes it is like building any other system the goal is like you build this once and then hopefully the cycles accelerate after that right and so now you could you could easily take a portion of that group and build something specifically you know for the nfl or for another entity much more quickly than you could could before but it's still a system that has to be built and put together and with it with outputs that that have to be managed a hundred percent and like uh perfect is on the first model we built i just binned on one day it's just like i have to get rid of this because it's just it's it's AI is moving so fast, the technological advancement.

21:43So I'd spent probably 10 of those 18 months working on a completely different model to the one we have now. But unless I would have tried to have built it the first way, I never would have got there to this way now. There is also sprouts that then come from the main one. So our product now, so if you subscribe up to 10B, is you get your own site with your own fans, no other fans. So you go, here is my five target audience category groups. Here's my sort of bands of super fans, sort of lower level fans and everything in between. We build you a custom version of that. So then in theory, if you're in the marketing department, you go, what are my casual fans thinking about X or Y?

22:26You just go on and just ask them a question. It doesn't need an agency to go away and give you a report. We're putting our tool in the hands of people that actually need it and want it. So yeah, it's a process. Yeah. Scott, I'm interested to get your view on the new PGA format that's being introduced in 2028, the postseason format. I think Brian Rollup has suggested that there's going to be a tour championship and a season championship. I think the big bet there from Rollup, who's the CEA of the tour, is that golf needs more drama, needs more jeopardy. And there's kind of match play format being introduced amongst other things.

23:10And I guess your big bet when you were at the tour is that golf needs more compelling stories. And that needs to be kind of delivered to fans through technology, ultimately. So I'm curious to know, you know, what your AI antenna was kind of doing when you sort of realized and saw that this postseason format is going to be introduced. You know, how are you, I guess, initially thinking about how the tool might use AI for a more complex and deeper format that's being introduced in golf?

23:40Scott Gutterman:Yeah, no, I'm pretty excited about it. I think it's a really interesting way to end the season by going in two weeks of match play. We're all still learning about the details of this, but two weeks of match play and then after going through this season with the end of the playoffs. I can tell you 20 years of the PGA Tour, some of our most exciting days were the Wednesdays of what used to be the Accenture match play that used to happen earlier in the year. Like Wednesdays, you have 64 players. They'd all come in. They'd be matched up. They'd be playing against each other. You're watching for elimination.

24:14Scott Gutterman:So Wednesday, Thursday, all of that was very exciting right up until the end where you We have two players going head to head. The use of AI in this particular case is going to be really interesting. One is I know that the teams have been, you know, they're going to be simulating what the outcomes of these events look like. Because at the tour, we were also very much using AI internally for a lot of initiatives, similar to what Ben was talking about. We simulated events, simulated events and formats that we were thinking about, simulated events and based on outcomes. And so the team there, the Shotlink team, the digital teams are all very strong at that, have gotten stronger than that.

24:50Scott Gutterman:And I'm sure that that's probably a tactic that they're looking at right now. The other part of it is that it's going to be a lot of excitement going on across. You described it earlier. It's 18 fields of play across 150, 200 acres. Match play, you're going to want to know what's happening. You need to know what's happening in whatever hole because it's going to be impacting the next matches coming up, either that day or the next day or the next week. And so I think the team will really approach it from how do we tell those stories quickly and easily and how do we personalize them so that you can follow your favorite players easily on the digital platforms that the PGA Tour offers or on any other platform that the PGA Tour wants to reach, right?

25:32Scott Gutterman:with that much content being generated all at once, whether it's video or stats or data. You want some of it going out to TikTok. You want some of it going to the old and operated platforms. And you want some of it going to, you know, the new, like we were talking about earlier, the new AI platforms that are being used now to look at what's going on live. And so, you know, probably when they get to 28, you know, we're going to be really interested to see what am I going to be asking ChatGPT to tell me what's going on on the PGA Tour and how does that affect consumption and personalization. And I think it's going to be even more interesting.

26:04Scott Gutterman:We're only actually it's tomorrow where Apple is going to be rolling out their updates for their hardware. We're going to be learning more about iOS 27. And there's some very big indications in iOS 27 that there's going to be some promising Siri integrations that are very much tied to what they're calling app intents that allow you allow Siri to access information in your apps and your platforms so that you can simply ask Siri for what for what's going on on site. So I think when they get to 28, the team's really going to be looking at how big an audience can we reach as quickly as possible with the information that they want.

26:39Scott Gutterman:And then also the idea of driving revenue opportunities from it as well. Yeah, absolutely. Chaps, I'm keen to talk examples, inspiration that you've seen probably over the last six to eight weeks, but maybe even more recently. Ben, I'm going to come to you first. when you sort of think of AI content generation in sport what has particularly stood out for you in recent weeks and why so there's there's a few sort of I don't know broadcast examples that have gone in the world I think just just talking about what we specialize in which is actually like the actual content creation for channels and that sort of stuff I think there's been some amazing advancements there's a company called Apptelligence who we met with a few times and they started to show me how how their model works and basically what they have created is the ability to create profiles front side back of players and do in theory whatever is then wanted with that said person right so perfect example is rodri joining barcelona as a announcement video ballon d 'or winner like the biggest thing he hadn't yet landed in Barcelona they were able to create a piece of content that featured him with a ticking clock sitting there just waiting I don't know if you if you've seen it or not it is not him and it is it is him but it's not him right they showed me exactly frame by frame how it's worked so therefore they had the ability to announce a story without some rubbish graphic that no one really cares about they could use content they could post it on tiktok they could have the video content at their fingertips telling a more interesting story to what scott's describing before but none of it was real but it looked really really real and there's a ton of different examples that that they've worked on we've now started working with that platform to to test things to show clients here's an idea and what it could look like but i think that how transfer ai sort of unfolded premier league transfer unfolded there was a lot of really bad versions of it right i think everyone probably saw some of that when it was done really well i think like in the barcelona example i think napoli did a few great examples fabrizio romano how he was doing some really clever stuff with kit swaps and it's effortless for these guys it takes seconds to make that type of content yeah but it's all powered by very smart back-end systems that are creating content now that sound shadows texture all of that so for me i always look at like the content itself and what that's being made to look like but there's tons of other stuff that i've i've seen i think i'll throw to scott for his better example than mine oh no no pressure there's two uh there's two that come to mind uh quickly for me one is i was watching the f1 race at monza uh this weekend and just watching kimmy antonelli come from 19th to win the win the race and i think he was the first italian to win the win there is in like 60 years or over 60 years and so like being able to have i'm a big fan of the f1 app and the f1 data uh extensions that they have on the app that you can subscribe to and i really like with a race like that watching that app watching the race come to life on that app and being able to see who's quickest in the turns and which sectors and everything like that.

30:08Scott Gutterman:That's always been something I'm always been a big fan of. And so that was an exciting, it was one of the most exciting races and I think in several years with F1. And so it was great to have all that different content available. The other one I'm going to go to real quick is SailGP, which continues to kind of build, I think, a lot of momentum through the content they create. And if you guys, I don't know if you've ever watched it, but SailGP, the AI and the graphics that they use on the broadcast, to show you the races as they're going is astounding. Because sailing, I watched a little bit of it before SailGP kind of emerged.

30:41Scott Gutterman:It was always very hard to tell who was winning what position and where on the course that they were on. And the work that the SailGP team has done on the broadcast with graphics and AI to kind of show you how that happens and how that content, how the race emerges is really fascinating to me. And it's also just made for like, it's a really good follow. Like I can follow now SailGP. And it's like a sport that I didn't think I was going to be, I would have never thought to follow. I have a good friend, Melissa Lawton, who's over there who heads up content there. And so they've done a great job on kind of continuing to do great content, but it's all through the use of these new tools.

31:20Scott Gutterman:And those are just two examples I really enjoy. very good a couple of things that have stood out to me lately um so firstly Real Madrid partnering with Adobe and I think that they partnered uh in 2025 so this is an extension to their partnership but basically allowing fans to co-create graphics um and insert themselves into sort of Real Madrid settings so with the players in a team photo you could insert yourself um insert a picture of yourself sorry lifting the trophy um it was kind of galactico theme so some of the settings are kind of in space it looked really cool and it was a great way for i guess madrid supporters to sort of immerse themselves with their favorite players um a couple of others uh so i noticed um a tool developed by uh neil horowitz i'm not sure if you have come across him chaps but he's um director of research and strategy at Greenfly and he's developed his own AI tool he also hosts his own sort of digital and social media themed podcast and he's developed his own tool called Skipper which essentially allows you to add your prompt and search back through I guess 13 years I think it is of archive podcast material and you can take whatever you're looking for so say uh i don't know you are interested in a certain um podcast guest that you know has been on there and you can say tell me everything that so and so said about this social strategy or tell me um all of the nfl executives that have appeared on this podcast and what they've said about um ownership in the league i thought that was a really sort of compelling example of somebody an individual launching their own sort of AI sort of tool for their own podcast that stood out to me, I think, in the last week on social.

33:12So hold on to you, Neil. That's very clever. Scott, I want to ask you about sort of this balance between experimentation with AI content generation and operationalization, which is a difficult word for me to say, obviously. um but yeah i guess how do you go from experimenting with ar to operationalizing it like is there something where you go we've done the tests is there something that you look for sorry where you go we've done the testing we feel like we're ready internally now's the time to actually try it and sort of service our fans or our customers yeah you know for us when we when i was at the tour

33:55Scott Gutterman:Or it was, you know, we had spent probably, you know, six months to a year doing proof of concepts and they were all good. And it was all about our staff learning closely with the AWS team on Bedrock and Claude and several other models, the AWS Nova models. But at some point, like you said, it's like, guys, we need to build something that fans will use or our internal staff will use. and the easiest tipping point is just like POCs are nice for learning, but we still have to provide value to the business. And so that is really what it comes down to, which is like, all right, I've done enough where I've learned enough.

34:32Scott Gutterman:How do we apply this to our business? And that's where we really started with, again, like we talked about the AI commentary that's on TorGast. Internally, we turned also on the broadcast technology side is can we automate the production of a par three broadcast that we do for some of our rights holders. And so how can we use it to control cameras and switch cameras based on a combination of our shot link data and a combination of knowing where the cameras are. In this par three, they're actually fixed cameras. So can we get them to zoom in? Because from shot link, we knew where the ball was going to land.

35:06Scott Gutterman:And so how could we actually do that and help that improve the cost and the quality of it? a quality. So those are two really good examples of like how we kind of made the leap forward into it. You know, at some point when they do every shot live at the Players Championship, they'll probably look at, you know, there's, you know, over a hundred cameras that are out there. They'll probably look at that production as well too. And how can you bring audio voice into that? And so it's again about value. We also began to find things where there was value for sponsorships. So AWS was good enough and came in and sponsored the AI commentary.

35:41Scott Gutterman:So we had some value there. The other thing that we did is we developed what was called one of our product managers came up with the idea of what was called the hot streak or what is called the hot streak at a PGA Tour event. It's a little... Chipotle, is that right? Yeah, I saw it. Yeah, Chipotle came in the sales teams. But there's a really good example of... It really anchored on two things. One is what can we... Number one, what can we do with the fans and get the fans to spend more time on our platforms? So you show somebody in the moment how a player is performing and that this is the best performing player on the course.

36:13Scott Gutterman:The people start, you know, fans start digging in on the apps and the websites to say, oh, how, why are they doing so well? Typically it would be their win probability would increase by a certain amount and they had a birdie streak or something to that effect. So that's when the little flame would go up there. And so we did that, put that up. We saw, I think it was a 15 or 20 % increase in scorecard consumption when we put that up there, which was fantastic. And then the sales teams, the revenue teams came along and said, hey, what can we do with this for our sponsors? And sponsors don't want to just be a part of spots and dots anymore.