1 00:00:00,160 --> 00:00:02,840 Speaker 1: Hey everyone, it's Robert and Joe here. Today we've got 2 00:00:02,880 --> 00:00:04,760 Speaker 1: something a little bit different to share with you. It 3 00:00:04,840 --> 00:00:08,039 Speaker 1: is a new season of the Smart Talks with IBM 4 00:00:08,160 --> 00:00:09,160 Speaker 1: podcast series. 5 00:00:09,640 --> 00:00:11,639 Speaker 2: Today we are witnessed to one of those rare moments 6 00:00:11,680 --> 00:00:14,400 Speaker 2: in history, the rise of an innovative technology with the 7 00:00:14,400 --> 00:00:18,680 Speaker 2: potential to radically transform business and society forever. The technology, 8 00:00:18,760 --> 00:00:22,200 Speaker 2: of course, is artificial intelligence, and it's the central focus 9 00:00:22,239 --> 00:00:24,840 Speaker 2: for this new season of Smart Talks with IBM. 10 00:00:25,320 --> 00:00:28,400 Speaker 1: Join hosts from your favorite Pushkin podcasts as they talk 11 00:00:28,480 --> 00:00:31,680 Speaker 1: with industry experts and leaders to explore how businesses can 12 00:00:31,720 --> 00:00:35,400 Speaker 1: integrate AI into their workflows and help drive real change 13 00:00:35,400 --> 00:00:38,240 Speaker 1: in this new era of AI. And of course, host 14 00:00:38,280 --> 00:00:40,479 Speaker 1: Malcolm Gladwell will be there to guide you through the 15 00:00:40,479 --> 00:00:42,640 Speaker 1: season and throw in his two cents as well. 16 00:00:43,120 --> 00:00:46,200 Speaker 2: Look out for new episodes of Smart Talks with IBM 17 00:00:46,400 --> 00:00:49,559 Speaker 2: every other week on the iHeartRadio app, Apple Podcasts, or 18 00:00:49,600 --> 00:00:53,360 Speaker 2: wherever you get your podcasts, and learn more at IBM 19 00:00:53,479 --> 00:00:55,480 Speaker 2: dot com slash smart talks. 20 00:00:57,160 --> 00:01:00,639 Speaker 3: Hello, Hello, Welcome to Smart Talks with IBM, a podcast 21 00:01:00,640 --> 00:01:05,920 Speaker 3: from Pushkin Industries, iHeartRadio and IBM. I'm Malcolm Gladwell. This 22 00:01:06,040 --> 00:01:09,160 Speaker 3: season we're diving back into the world of artificial intelligence, 23 00:01:09,560 --> 00:01:12,400 Speaker 3: but with a focus on the powerful concept of open 24 00:01:12,760 --> 00:01:17,800 Speaker 3: its possibilities, implications, and misconceptions. We'll look at openness from 25 00:01:17,800 --> 00:01:20,760 Speaker 3: a variety of angles and explore how the concept is 26 00:01:20,800 --> 00:01:24,680 Speaker 3: already reshaping industries, ways of doing business, and our very 27 00:01:24,720 --> 00:01:29,960 Speaker 3: notion of what's possible. I'm particularly excited for today's guest, 28 00:01:30,319 --> 00:01:34,440 Speaker 3: Brian Ryerson. He's Senior Director of Digital Strategy at the 29 00:01:34,560 --> 00:01:38,080 Speaker 3: US Tennis Association, helping to oversee one of the most 30 00:01:38,280 --> 00:01:41,520 Speaker 3: iconic events in the world of sports, the US Open. 31 00:01:42,480 --> 00:01:45,840 Speaker 3: Brian sat down with Pushkin's own Jacob Goldstein, host of 32 00:01:45,880 --> 00:01:50,320 Speaker 3: the podcast What's Your Problem. A veteran business journalist, Jacob 33 00:01:50,320 --> 00:01:53,240 Speaker 3: has reported for The Wall Street Journal, the Miami Herald, 34 00:01:53,440 --> 00:01:57,920 Speaker 3: and was a longtime host of the NPR program Planet Money. 35 00:01:58,280 --> 00:02:01,600 Speaker 3: IBM has been the official technology partner of the US 36 00:02:01,640 --> 00:02:05,800 Speaker 3: Tennis Association for more than thirty years, and the more 37 00:02:05,840 --> 00:02:10,400 Speaker 3: recent evolution into generative AI has enhanced the world class 38 00:02:10,440 --> 00:02:14,840 Speaker 3: digital experiences that help more than fifteen million fans from 39 00:02:14,919 --> 00:02:18,480 Speaker 3: all over the world enjoy the US Open Tennis Championships. 40 00:02:18,919 --> 00:02:22,880 Speaker 3: In this episode, we will explore how generative AI is 41 00:02:22,919 --> 00:02:27,560 Speaker 3: being used to generate match insights, spoken commentary for match highlights, 42 00:02:27,840 --> 00:02:31,640 Speaker 3: and postmatch summaries at scale for fans to enjoy through 43 00:02:31,680 --> 00:02:35,040 Speaker 3: the US Open app and website. We'll explore how these 44 00:02:35,080 --> 00:02:39,520 Speaker 3: AI solutions enable the editorial team to cover more of 45 00:02:39,560 --> 00:02:43,760 Speaker 3: the tournament than ever before, bringing fans even closer to 46 00:02:43,840 --> 00:02:46,160 Speaker 3: the game they love, and will learn more about one 47 00:02:46,200 --> 00:02:50,360 Speaker 3: of the engines behind this AI powered content creation, a 48 00:02:50,440 --> 00:02:54,160 Speaker 3: large language model from the ib M Granite family, which 49 00:02:54,200 --> 00:02:59,560 Speaker 3: is trained and maintained using the watsonex AI and data platform. Okay, 50 00:03:00,360 --> 00:03:01,040 Speaker 3: let's dive in. 51 00:03:02,320 --> 00:03:03,400 Speaker 4: Brian, Welcome to the show. 52 00:03:03,639 --> 00:03:05,079 Speaker 5: Thanks for having me. I'm excited to be here. 53 00:03:05,480 --> 00:03:07,000 Speaker 4: Can you say your name and your job? 54 00:03:07,400 --> 00:03:10,720 Speaker 5: Yeah, I'm Brian Ryerson. I'm Senior director of Digital Strategy 55 00:03:10,720 --> 00:03:11,440 Speaker 5: at the USTA. 56 00:03:11,800 --> 00:03:14,160 Speaker 4: Some question, what's the USTA. 57 00:03:13,960 --> 00:03:16,080 Speaker 5: The US Tennis Association. 58 00:03:16,440 --> 00:03:19,000 Speaker 4: And tell me about the USTA, Like, what is it? 59 00:03:19,440 --> 00:03:22,000 Speaker 5: Yeah? So the USTA is the governing body of tennis 60 00:03:22,000 --> 00:03:25,360 Speaker 5: in the US. Or mission is to grow the sport 61 00:03:25,400 --> 00:03:28,720 Speaker 5: of tennis across the US at all levels. Really, I 62 00:03:28,760 --> 00:03:30,840 Speaker 5: would say we're more like a health and wellness company 63 00:03:30,880 --> 00:03:33,480 Speaker 5: where tennis is the means to health and wellness. And 64 00:03:33,520 --> 00:03:36,040 Speaker 5: then the US Open is kind of our tenth pole 65 00:03:36,080 --> 00:03:38,960 Speaker 5: event that happens everyear and Flushing Meadows and is really 66 00:03:38,960 --> 00:03:41,040 Speaker 5: our chance to showcase the support of tennis at its 67 00:03:41,120 --> 00:03:43,240 Speaker 5: highest level to fans all around the world. 68 00:03:43,320 --> 00:03:45,480 Speaker 4: Yeah, I mean the US Open. I assume most people 69 00:03:45,560 --> 00:03:47,680 Speaker 4: know this, but it's Grand Slam. It's one of the 70 00:03:47,760 --> 00:03:50,800 Speaker 4: what four biggest tennis tournaments in the world. 71 00:03:50,920 --> 00:03:54,160 Speaker 5: Yes, yeah, every year, we especially the past couple of years, 72 00:03:54,160 --> 00:03:57,480 Speaker 5: we've seen immense growth and you know, we are very 73 00:03:57,560 --> 00:03:59,480 Speaker 5: hopeful this year and our big goals that have over 74 00:03:59,480 --> 00:04:01,680 Speaker 5: a million fires on site during the three week window 75 00:04:01,720 --> 00:04:04,000 Speaker 5: this year, So it's an amazing event. I always say 76 00:04:04,040 --> 00:04:06,320 Speaker 5: it's a food and wine festival where tennis is the 77 00:04:06,360 --> 00:04:09,760 Speaker 5: main attraction and it's a really fun, unique atmosphere. 78 00:04:10,440 --> 00:04:12,640 Speaker 4: How did you get into the tennis business. 79 00:04:12,440 --> 00:04:15,160 Speaker 5: It's a great question. It's not where I thought i'd 80 00:04:15,240 --> 00:04:17,640 Speaker 5: end up for especially being there for fourteen years. So 81 00:04:17,839 --> 00:04:21,200 Speaker 5: I was a marketing and technology major in school, and 82 00:04:21,880 --> 00:04:24,480 Speaker 5: I also played college lacrosse and sports was always a 83 00:04:24,480 --> 00:04:26,160 Speaker 5: big part of my life and always wanted to be 84 00:04:26,200 --> 00:04:28,600 Speaker 5: in the sports and entertainment world. I'm here from the 85 00:04:28,640 --> 00:04:30,200 Speaker 5: New York area. This is where I grew up so 86 00:04:30,240 --> 00:04:32,680 Speaker 5: I moved back home and had a few friends who 87 00:04:32,720 --> 00:04:35,320 Speaker 5: worked there, and I started out more on the number 88 00:04:35,400 --> 00:04:37,640 Speaker 5: side of things and really digital analytics. It was really 89 00:04:37,680 --> 00:04:41,240 Speaker 5: the start of when Facebook and Twitter is just starting 90 00:04:41,240 --> 00:04:44,080 Speaker 5: and digital marketing and all of that. And you know, 91 00:04:44,160 --> 00:04:46,320 Speaker 5: I went to my first year so Open not really 92 00:04:46,400 --> 00:04:49,680 Speaker 5: knowing what to expect, and again, I think the atmosphere 93 00:04:49,760 --> 00:04:52,120 Speaker 5: kind of captivated me and hooked me in. And I've 94 00:04:52,120 --> 00:04:53,400 Speaker 5: been there now fourteen years. 95 00:04:53,800 --> 00:04:58,000 Speaker 4: And so your title is Digital Director. What does that mean? 96 00:04:58,080 --> 00:04:58,719 Speaker 4: What's your job? 97 00:04:58,920 --> 00:05:02,360 Speaker 5: Yeah, so it's interesting one because it's tough to explain 98 00:05:02,400 --> 00:05:04,800 Speaker 5: to folks who are not in the weeds on all 99 00:05:04,880 --> 00:05:07,080 Speaker 5: things US open or even in the sports world. But 100 00:05:07,200 --> 00:05:10,719 Speaker 5: really I oversee all of our consumer facing digital property. 101 00:05:10,880 --> 00:05:13,359 Speaker 5: So that's the us open dot org, our website built 102 00:05:13,360 --> 00:05:16,200 Speaker 5: by IBM, as well as our mobile app. I oversee 103 00:05:16,240 --> 00:05:21,320 Speaker 5: our content strategy, our sponsorship integrations. So really anything consumer 104 00:05:21,360 --> 00:05:24,400 Speaker 5: facing that happens on the web is under my purview, 105 00:05:24,680 --> 00:05:27,479 Speaker 5: even some of our new platform extensions and gaming and 106 00:05:27,520 --> 00:05:30,760 Speaker 5: things like that. Anything that you can physically interact with 107 00:05:31,120 --> 00:05:32,560 Speaker 5: is kind of under my purview. 108 00:05:33,279 --> 00:05:37,160 Speaker 4: And so you've been there now for fourteen ISSU years, 109 00:05:37,320 --> 00:05:40,960 Speaker 4: which in the digital world is a long time. How 110 00:05:41,000 --> 00:05:45,320 Speaker 4: has that sort of digital experience of sports changed over 111 00:05:45,360 --> 00:05:45,839 Speaker 4: that time. 112 00:05:46,080 --> 00:05:49,640 Speaker 5: Yeah, it's obviously grown digital now, is what we say 113 00:05:49,680 --> 00:05:51,800 Speaker 5: and what my team says. It's the number one way 114 00:05:51,839 --> 00:05:54,479 Speaker 5: to engage with fans that can't make it to the event, 115 00:05:54,680 --> 00:05:56,160 Speaker 5: as well as those fans who are at the event, 116 00:05:56,200 --> 00:05:58,000 Speaker 5: and how to enrich their stay. So it's really kind 117 00:05:58,040 --> 00:06:02,760 Speaker 5: of you're tackling multiple personas. It's the international fan who's 118 00:06:03,040 --> 00:06:05,200 Speaker 5: staying up late to watch in other countries, to the 119 00:06:05,720 --> 00:06:07,919 Speaker 5: fan here who's maybe watching on broadcasts, and we go 120 00:06:07,920 --> 00:06:10,520 Speaker 5: in a company and enrich that broadcast with new stats 121 00:06:10,560 --> 00:06:12,680 Speaker 5: and insights. To the on site fan who bought a 122 00:06:12,720 --> 00:06:15,000 Speaker 5: ticket and maybe doesn't know what match is happening on 123 00:06:15,040 --> 00:06:18,040 Speaker 5: what court. We do have twenty plus courts happening at 124 00:06:18,040 --> 00:06:20,279 Speaker 5: a time, with all different matches, So we really try 125 00:06:20,320 --> 00:06:23,279 Speaker 5: to help all fans navigate the US Open the best 126 00:06:23,279 --> 00:06:23,960 Speaker 5: way possible. 127 00:06:24,320 --> 00:06:26,480 Speaker 4: And so, like, what are some of the sort of 128 00:06:26,480 --> 00:06:28,479 Speaker 4: problems you're trying to solve. What are some of the 129 00:06:28,480 --> 00:06:29,800 Speaker 4: hard things about your job? 130 00:06:30,120 --> 00:06:34,240 Speaker 5: Yeah, obviously technology changes at a rapid pace, right, So 131 00:06:34,279 --> 00:06:36,159 Speaker 5: I think part of it is how do we stay 132 00:06:36,200 --> 00:06:38,320 Speaker 5: on the forefront of that, and how do we do 133 00:06:38,360 --> 00:06:40,520 Speaker 5: that in the best way and make the best fan 134 00:06:40,600 --> 00:06:44,560 Speaker 5: experiences possible and the best user experiences possible. That's always 135 00:06:44,760 --> 00:06:47,680 Speaker 5: kind of driving factor number one. Then number two, it's 136 00:06:48,120 --> 00:06:51,240 Speaker 5: understanding and listening to our fans and what kind of 137 00:06:51,279 --> 00:06:54,279 Speaker 5: content they want. You hear me talk a lot about storytelling. 138 00:06:54,279 --> 00:06:56,239 Speaker 5: I feel like there's a lot of storytelling that happens 139 00:06:56,240 --> 00:06:58,160 Speaker 5: around the US open that we really want to bring 140 00:06:58,200 --> 00:07:01,880 Speaker 5: to fans. And that can be as simple as storytelling 141 00:07:01,920 --> 00:07:05,000 Speaker 5: of what's happening today and what you should be watching too. 142 00:07:05,520 --> 00:07:08,200 Speaker 5: Maybe it's your favorite players and what's going on behind 143 00:07:08,240 --> 00:07:11,880 Speaker 5: the scenes with them, to even introducing I want to say, 144 00:07:11,880 --> 00:07:15,200 Speaker 5: the casual fans to who they should be watching, why 145 00:07:15,200 --> 00:07:17,720 Speaker 5: they should follow certain players, and more bringing that player's 146 00:07:17,760 --> 00:07:18,640 Speaker 5: story to life. 147 00:07:19,240 --> 00:07:19,480 Speaker 2: Yeah. 148 00:07:19,520 --> 00:07:21,800 Speaker 4: I mean, I feel like almost the whole point of 149 00:07:21,800 --> 00:07:25,080 Speaker 4: sports is to like create stories for us to follow, right, 150 00:07:25,160 --> 00:07:28,720 Speaker 4: Like they're engineered to be stories. It's exactly, this thing 151 00:07:28,800 --> 00:07:30,920 Speaker 4: is happening in front of you and there are two 152 00:07:31,480 --> 00:07:34,480 Speaker 4: antagonists and the stakes are high, and you don't know 153 00:07:34,480 --> 00:07:37,040 Speaker 4: how it's gonna end, Like it's built to be a story. 154 00:07:37,280 --> 00:07:39,520 Speaker 5: Yeah, And that's the main challenge of the job is 155 00:07:39,560 --> 00:07:42,320 Speaker 5: you can plan, plan, plan, but once you get on 156 00:07:42,480 --> 00:07:44,400 Speaker 5: two players on court and you don't know what that 157 00:07:44,440 --> 00:07:46,520 Speaker 5: outcome is going to be, it's now sitting and waiting 158 00:07:46,560 --> 00:07:49,000 Speaker 5: and watching and you become a fan yourself. And then 159 00:07:49,000 --> 00:07:52,440 Speaker 5: it's how do you really captivate that story and how 160 00:07:52,440 --> 00:07:54,720 Speaker 5: do you narrate it and how do you like translate 161 00:07:54,840 --> 00:07:55,720 Speaker 5: up to fans. 162 00:07:55,600 --> 00:07:57,000 Speaker 4: And it's like you kind of have to do it 163 00:07:57,040 --> 00:07:59,080 Speaker 4: in real time, right, Like the whole point of sports 164 00:07:59,120 --> 00:08:00,120 Speaker 4: is you don't know what's gonna. 165 00:07:59,920 --> 00:08:03,000 Speaker 5: Happen exactly, and that's the excitement. And it's also there's 166 00:08:03,000 --> 00:08:04,880 Speaker 5: so many different types of fans. You know, there's the 167 00:08:05,480 --> 00:08:07,560 Speaker 5: fans who want a lot of enriched data and their 168 00:08:07,640 --> 00:08:09,800 Speaker 5: tennis nerds for lack of better of saying it, and 169 00:08:09,840 --> 00:08:12,720 Speaker 5: that they really want to dive deep into the intricacies 170 00:08:12,720 --> 00:08:15,200 Speaker 5: of the game, versus the casual fan who maybe just 171 00:08:15,240 --> 00:08:17,280 Speaker 5: wants more of this high level storyline of what does 172 00:08:17,320 --> 00:08:20,720 Speaker 5: this mean? Why is it important? So it's really trying 173 00:08:20,720 --> 00:08:23,040 Speaker 5: to figure out how to deliver that at scale and 174 00:08:23,320 --> 00:08:25,880 Speaker 5: really help fans get what they're looking for and the 175 00:08:25,880 --> 00:08:27,200 Speaker 5: type of content they're looking for. 176 00:08:27,520 --> 00:08:30,720 Speaker 4: So are there specific examples of you know how fan 177 00:08:30,800 --> 00:08:35,360 Speaker 4: feedback has led to specific features digital features you build. 178 00:08:35,720 --> 00:08:38,840 Speaker 4: Are there, like particularly popular features you've come up with, Like, 179 00:08:38,840 --> 00:08:40,559 Speaker 4: what are some specifics. 180 00:08:40,000 --> 00:08:42,520 Speaker 5: Yeah, some low hanging fruit type things that came from 181 00:08:42,600 --> 00:08:45,960 Speaker 5: fan feedback. Is simple things sometimes like managing time zones 182 00:08:45,960 --> 00:08:47,319 Speaker 5: and when matches start. 183 00:08:47,280 --> 00:08:50,520 Speaker 4: A persistent problem where those of us can work across 184 00:08:50,559 --> 00:08:52,000 Speaker 4: times exactly. 185 00:08:52,280 --> 00:08:55,080 Speaker 5: And we do have, like I've mentioned, twenty plus courts 186 00:08:55,080 --> 00:08:57,240 Speaker 5: happening at a time, So it's a lot to follow, 187 00:08:57,280 --> 00:08:59,720 Speaker 5: and how do you translate that to a fan, whether 188 00:08:59,720 --> 00:09:02,360 Speaker 5: it's to their native language or to their time zone 189 00:09:02,400 --> 00:09:04,600 Speaker 5: or things like that. So that's one thing that came 190 00:09:04,640 --> 00:09:07,560 Speaker 5: through fan feedback, and another one a three to five 191 00:09:07,640 --> 00:09:10,280 Speaker 5: hour match, especially when you're having twenty plus of them 192 00:09:10,280 --> 00:09:12,439 Speaker 5: happening at a time, is there's too much for one 193 00:09:12,480 --> 00:09:15,920 Speaker 5: person to follow. So how do you start from an 194 00:09:16,040 --> 00:09:19,400 Speaker 5: editorial perspective really helping with that storytelling and guiding a 195 00:09:19,440 --> 00:09:22,040 Speaker 5: fan to like, all right, whether there's an upset about 196 00:09:22,040 --> 00:09:25,960 Speaker 5: to happen, or here's your matches to watch, or even 197 00:09:26,000 --> 00:09:27,760 Speaker 5: some of the predictions we're starting to put in, is 198 00:09:27,760 --> 00:09:29,760 Speaker 5: we really want to guide the fan before a match, 199 00:09:30,120 --> 00:09:32,680 Speaker 5: here's where you should tune in to even after a 200 00:09:32,679 --> 00:09:35,959 Speaker 5: match of here's what's happened, Here's what's important, and we're 201 00:09:36,000 --> 00:09:37,559 Speaker 5: really excited with some of the features we've built in 202 00:09:37,600 --> 00:09:40,200 Speaker 5: the last few years that I would say really helps 203 00:09:40,280 --> 00:09:41,920 Speaker 5: us do that at more scale than what we were 204 00:09:41,960 --> 00:09:44,240 Speaker 5: able to do with just writers following a match and 205 00:09:44,320 --> 00:09:45,520 Speaker 5: covering every single match. 206 00:09:45,760 --> 00:09:48,520 Speaker 4: Huh. So I want to talk a little bit about 207 00:09:48,880 --> 00:09:54,120 Speaker 4: the partnership between IBM and the USTA, Like, just tell 208 00:09:54,160 --> 00:09:55,679 Speaker 4: me about the work you do together. 209 00:09:56,080 --> 00:09:59,320 Speaker 5: So IBM is our official digital and technology partner and 210 00:09:59,360 --> 00:10:02,720 Speaker 5: innovation the US Open they predate me. It's a thirty 211 00:10:02,800 --> 00:10:05,600 Speaker 5: year partnership and it truly as a partnership. So I 212 00:10:05,720 --> 00:10:08,720 Speaker 5: view the IBM consulting team as an extension of my 213 00:10:09,440 --> 00:10:12,880 Speaker 5: USTA team, So we work with them year round. They design, 214 00:10:13,200 --> 00:10:17,040 Speaker 5: develop and deliver the digital properties. They help us provide 215 00:10:17,040 --> 00:10:19,760 Speaker 5: the tools to create content to do things at scale, 216 00:10:19,800 --> 00:10:22,520 Speaker 5: They help us from stats and information and really help 217 00:10:22,600 --> 00:10:24,960 Speaker 5: us push from an innovation standpoint to make sure that 218 00:10:25,000 --> 00:10:27,200 Speaker 5: we are staying on that cutting edge of technology. So 219 00:10:27,840 --> 00:10:31,560 Speaker 5: I would truly say it's much more than a sponsorship, 220 00:10:31,559 --> 00:10:34,800 Speaker 5: where it's truly a partnership to deliver that fan experience. 221 00:10:34,760 --> 00:10:37,400 Speaker 4: And so What are some of the specific things that 222 00:10:37,640 --> 00:10:38,800 Speaker 4: you have done with IBM. 223 00:10:39,080 --> 00:10:43,240 Speaker 5: Yeah, so, I mean there's countless ones to talk through. Obviously, 224 00:10:43,280 --> 00:10:45,880 Speaker 5: they thirty years ago. They helped us build our first 225 00:10:45,920 --> 00:10:49,400 Speaker 5: website and it's kind of grown from there over the 226 00:10:49,559 --> 00:10:51,400 Speaker 5: past few years. I would say, I think it was 227 00:10:51,400 --> 00:10:53,959 Speaker 5: twenty eighteen as we started AI Highlights, So that was 228 00:10:54,000 --> 00:10:57,720 Speaker 5: really when we were able to have all twenty matches 229 00:10:57,760 --> 00:11:00,000 Speaker 5: going at a single time. We were able to quickly 230 00:11:00,200 --> 00:11:03,800 Speaker 5: deliver succinct highlights to fans to our digital platform so 231 00:11:03,840 --> 00:11:06,080 Speaker 5: they could see highlights for every single core. 232 00:11:06,720 --> 00:11:09,679 Speaker 4: Is that video highlights? Is that tech summaries? What does 233 00:11:09,720 --> 00:11:10,160 Speaker 4: that mean? 234 00:11:10,640 --> 00:11:13,160 Speaker 5: At the time, it was video highlights, Okay, So it 235 00:11:13,200 --> 00:11:16,040 Speaker 5: was really taking that three to five hour match, let's say, 236 00:11:16,080 --> 00:11:18,559 Speaker 5: and cut it down to a three minute highlight that 237 00:11:18,720 --> 00:11:21,080 Speaker 5: could show up within moments after a match, ending to 238 00:11:21,160 --> 00:11:23,640 Speaker 5: our website and our mobile app, so fans could see 239 00:11:23,640 --> 00:11:25,280 Speaker 5: that all around the world and really kind of get 240 00:11:25,320 --> 00:11:28,480 Speaker 5: that three minute overview what happened in a match, and. 241 00:11:28,600 --> 00:11:31,720 Speaker 4: Was that AI enabled? Was AI a piece of how 242 00:11:31,720 --> 00:11:32,199 Speaker 4: to do that? 243 00:11:32,320 --> 00:11:34,959 Speaker 5: It was? It was probably our first foray into AI. 244 00:11:35,600 --> 00:11:42,400 Speaker 4: Back twenty eighteen is relatively early, Yeah, exactly, for tennis exactly. 245 00:11:42,480 --> 00:11:45,439 Speaker 5: Yeah. It really, I want to say, opened up our 246 00:11:45,480 --> 00:11:49,439 Speaker 5: ability to one again storytell but attract new fans too. 247 00:11:49,559 --> 00:11:51,800 Speaker 5: Is video has actually been our number one growth area 248 00:11:51,880 --> 00:11:54,040 Speaker 5: since twenty eighteen, and I think a lot of that 249 00:11:54,080 --> 00:11:55,800 Speaker 5: has to do with the scale of how we deliver 250 00:11:55,880 --> 00:11:56,800 Speaker 5: that content. 251 00:11:56,720 --> 00:11:59,640 Speaker 4: Using AI and being able to deliver this sort of 252 00:11:59,720 --> 00:12:02,920 Speaker 4: video highlight reels at scale. 253 00:12:02,760 --> 00:12:05,320 Speaker 5: Yeah, and do it quickly. Right. We've always had highlights, 254 00:12:05,360 --> 00:12:07,319 Speaker 5: but it was a manual process where you had a 255 00:12:07,880 --> 00:12:11,079 Speaker 5: video that or cutting through you know, a three hour match, 256 00:12:11,200 --> 00:12:13,480 Speaker 5: selecting the right scene, stitching together, it would have to 257 00:12:13,520 --> 00:12:16,720 Speaker 5: get voiced over, et cetera. We really have used AI 258 00:12:16,840 --> 00:12:18,400 Speaker 5: to make it, i want to say, much more efficient 259 00:12:18,520 --> 00:12:21,520 Speaker 5: and speed up that process and deliver it more quickly 260 00:12:21,559 --> 00:12:22,160 Speaker 5: to our fans. 261 00:12:22,520 --> 00:12:24,439 Speaker 4: I mean, it would be a bummer to get scooped 262 00:12:24,600 --> 00:12:27,679 Speaker 4: by whatever NBC News or Yes Pen or whatever. I'm 263 00:12:27,679 --> 00:12:29,480 Speaker 4: sure there are all your partners and you love that most. 264 00:12:29,960 --> 00:12:32,240 Speaker 4: Obviously you want to have the video first, right, it's 265 00:12:32,280 --> 00:12:32,760 Speaker 4: your match. 266 00:12:33,000 --> 00:12:35,200 Speaker 5: Yeah, And I think it's also important to us as 267 00:12:35,559 --> 00:12:39,719 Speaker 5: being the USTA is ensuring that it's not just you know, 268 00:12:39,880 --> 00:12:43,440 Speaker 5: the main marquee players, that every player and all those 269 00:12:43,480 --> 00:12:46,960 Speaker 5: storylines and that whether it's you know, the main singles 270 00:12:47,040 --> 00:12:49,679 Speaker 5: draw to our mixed doubles, et cetera. They all need 271 00:12:49,760 --> 00:12:52,200 Speaker 5: highlights and they all have their own stories to tell, 272 00:12:52,200 --> 00:12:53,960 Speaker 5: and how do we do that at scale? It was 273 00:12:53,960 --> 00:12:56,400 Speaker 5: something that before we had that product was not something 274 00:12:56,440 --> 00:12:57,160 Speaker 5: you were able to do. 275 00:12:57,600 --> 00:13:01,040 Speaker 4: Great, So let's let's talk in some more detail about 276 00:13:01,080 --> 00:13:04,480 Speaker 4: what you're working on. Let's start with the app. Tell 277 00:13:04,480 --> 00:13:07,520 Speaker 4: me about the us Open app and the Companion website. 278 00:13:07,640 --> 00:13:10,120 Speaker 5: Yeah, so I'll start with the app, and I feel 279 00:13:10,200 --> 00:13:13,600 Speaker 5: like they serve similar needs, but they're a little different 280 00:13:13,640 --> 00:13:16,720 Speaker 5: in their own respective manners. Is the app. Everybody has 281 00:13:16,760 --> 00:13:18,640 Speaker 5: a phone in their hands at this point. The app 282 00:13:18,720 --> 00:13:20,880 Speaker 5: is kind of their guide to when I say a 283 00:13:20,920 --> 00:13:23,600 Speaker 5: million fans on site, we view the app as we 284 00:13:23,640 --> 00:13:26,160 Speaker 5: want that to be their on site guide, and Companion 285 00:13:26,400 --> 00:13:27,040 Speaker 5: a million. 286 00:13:27,160 --> 00:13:29,880 Speaker 4: Let's just pause on a million fans on site, right, 287 00:13:29,880 --> 00:13:33,640 Speaker 4: because like a big professional whatever, an NFL game or 288 00:13:33,640 --> 00:13:37,120 Speaker 4: something that's like one hundred thousand, this is ten x that. 289 00:13:37,480 --> 00:13:40,520 Speaker 5: Yeah, and a three week window and a very succinct, 290 00:13:40,679 --> 00:13:45,080 Speaker 5: tight action packed window. There's a lot of action logistics. 291 00:13:45,160 --> 00:13:46,880 Speaker 4: Okay, so keep going. 292 00:13:46,960 --> 00:13:49,800 Speaker 5: So the app, you know, whether it's finding the schedules, 293 00:13:49,840 --> 00:13:52,760 Speaker 5: the live scores, what's happening on court. That's really the 294 00:13:52,760 --> 00:13:55,880 Speaker 5: focus point of the app, and what we're really focused 295 00:13:55,880 --> 00:13:58,120 Speaker 5: on this year is how do we build in some 296 00:13:58,160 --> 00:14:00,839 Speaker 5: of those map summaries into the app, into our slam 297 00:14:00,880 --> 00:14:04,200 Speaker 5: Tracker experience. So again, before match, that kind of match 298 00:14:04,240 --> 00:14:06,480 Speaker 5: preview of here's maybe if you have a ticket, here's 299 00:14:06,520 --> 00:14:09,360 Speaker 5: what to expect, here's you know are likely to win, 300 00:14:09,440 --> 00:14:11,920 Speaker 5: who we are predicting, so you can kind of get 301 00:14:11,960 --> 00:14:15,320 Speaker 5: some information heading in, and then after the match it's 302 00:14:15,320 --> 00:14:18,559 Speaker 5: more of what just happened, what it means for the 303 00:14:18,800 --> 00:14:21,560 Speaker 5: rest of the draw, who they're playing next, is this 304 00:14:21,640 --> 00:14:23,800 Speaker 5: the first time this has happened, et cetera, and really 305 00:14:23,920 --> 00:14:26,960 Speaker 5: enriching that experience as well. So the app is one 306 00:14:27,040 --> 00:14:29,360 Speaker 5: your guide to what you should be watching, but also 307 00:14:29,400 --> 00:14:31,960 Speaker 5: then giving you that insights and context of what's happening 308 00:14:32,000 --> 00:14:32,960 Speaker 5: on that court as you're. 309 00:14:32,880 --> 00:14:36,480 Speaker 4: Watching, like the commentator in your pocket exactly. So you 310 00:14:36,640 --> 00:14:39,400 Speaker 4: used a phrase in there as if I already knew it, 311 00:14:40,280 --> 00:14:41,880 Speaker 4: and I love the phrase, but I want you to 312 00:14:41,880 --> 00:14:44,760 Speaker 4: talk more about it. That phrase is slam Tracker. 313 00:14:45,000 --> 00:14:50,720 Speaker 5: Yes, So slam Tracker is our long standing live scores 314 00:14:50,760 --> 00:14:53,040 Speaker 5: I want to say Match Center. It is, okay, where 315 00:14:53,240 --> 00:14:55,960 Speaker 5: every single data point for every single match lives, and 316 00:14:56,000 --> 00:14:59,280 Speaker 5: it really it helps showcase what's happening to match. I say, 317 00:14:59,280 --> 00:15:02,120 Speaker 5: it's our brock Cast companions. If you're watching live, it's 318 00:15:02,120 --> 00:15:04,640 Speaker 5: our in stadium companion. And it's also the best thing 319 00:15:04,720 --> 00:15:06,720 Speaker 5: to have if you aren't able to watch. 320 00:15:06,840 --> 00:15:08,600 Speaker 4: And so, like, I'm on the app and there's a 321 00:15:08,680 --> 00:15:11,520 Speaker 4: thing called slam Tracker, and it like taps slam Tracker. 322 00:15:11,560 --> 00:15:13,320 Speaker 4: What do I see on my phone when I tap 323 00:15:13,360 --> 00:15:16,720 Speaker 4: slam Tracker? You know, midday when the tournament's happening. 324 00:15:16,760 --> 00:15:18,640 Speaker 5: So before match, that's where you get a lot of 325 00:15:18,640 --> 00:15:20,840 Speaker 5: pre match content. That's where those live kind of our 326 00:15:20,880 --> 00:15:24,240 Speaker 5: predictions are likelihood to win lives within that So likelihood 327 00:15:24,280 --> 00:15:26,840 Speaker 5: to win essentially pulls in a bunch of data points. 328 00:15:26,880 --> 00:15:30,000 Speaker 5: So pass matches, how many times these players have played 329 00:15:30,000 --> 00:15:32,680 Speaker 5: each other against each other, even some punditry and other 330 00:15:32,720 --> 00:15:35,520 Speaker 5: written articles that maybe our editorial team put out and 331 00:15:35,560 --> 00:15:37,920 Speaker 5: really kind of puts a prediction out there. 332 00:15:38,040 --> 00:15:40,160 Speaker 4: And so it's just a percentage chance. 333 00:15:40,080 --> 00:15:43,160 Speaker 5: Yes exactly, but it uses millions of data points to 334 00:15:43,200 --> 00:15:45,480 Speaker 5: come up with that. Yes, So it really helps you 335 00:15:45,560 --> 00:15:48,520 Speaker 5: kind of understand what you're getting into for that match. 336 00:15:48,920 --> 00:15:51,640 Speaker 5: During a live match, it is every single point, so 337 00:15:52,000 --> 00:15:55,200 Speaker 5: point by point scoring as well as in depth analysis 338 00:15:55,200 --> 00:15:58,080 Speaker 5: in point commentary where also this year have a live 339 00:15:58,200 --> 00:16:02,080 Speaker 5: visualization that accompanies that will really help bring the match together. 340 00:16:02,520 --> 00:16:04,240 Speaker 5: And what I mean by that is it uses our 341 00:16:04,280 --> 00:16:07,880 Speaker 5: ball tracking technology to really showcase the match in near 342 00:16:07,920 --> 00:16:10,720 Speaker 5: real time, so within seconds delay of where the ball's 343 00:16:10,920 --> 00:16:13,360 Speaker 5: being hit, where the players are, and really bring a 344 00:16:13,480 --> 00:16:16,240 Speaker 5: visualization to life and layered stats and data on top 345 00:16:16,280 --> 00:16:16,400 Speaker 5: of it. 346 00:16:16,520 --> 00:16:16,640 Speaker 2: Huh. 347 00:16:16,960 --> 00:16:18,720 Speaker 4: Is that sort of like when I'm watching a match 348 00:16:18,760 --> 00:16:21,360 Speaker 4: on TV and there's like a close call as the 349 00:16:21,360 --> 00:16:23,120 Speaker 4: ball in or out and they do that thing where 350 00:16:23,120 --> 00:16:25,320 Speaker 4: they kind of show a sort of video game version 351 00:16:25,360 --> 00:16:27,160 Speaker 4: of where the ball landed. Does it look like that? 352 00:16:27,440 --> 00:16:30,120 Speaker 5: It's like that before every single shot, So it's not 353 00:16:30,280 --> 00:16:32,720 Speaker 5: just those close ones. It's our first foray to bring 354 00:16:32,760 --> 00:16:33,760 Speaker 5: that match to life. 355 00:16:34,280 --> 00:16:36,120 Speaker 4: Huh. And so what do I see on that kind 356 00:16:36,120 --> 00:16:38,320 Speaker 4: of view that I don't see from whatever watching the video? 357 00:16:38,440 --> 00:16:41,080 Speaker 5: Yeah, So one you'll be able just to see more 358 00:16:41,120 --> 00:16:44,280 Speaker 5: of the ball trajectory and where the ball is being hit, 359 00:16:44,320 --> 00:16:46,640 Speaker 5: but then you can also start layering things in stats 360 00:16:46,640 --> 00:16:49,120 Speaker 5: and insights on top of that, so how many times 361 00:16:49,160 --> 00:16:52,359 Speaker 5: has player A hit the ball on a certain baseline, 362 00:16:52,360 --> 00:16:55,320 Speaker 5: how fast are they hitting it, maybe their serve percentage 363 00:16:55,320 --> 00:16:57,160 Speaker 5: and a certain side of the court, et cetera. So 364 00:16:57,200 --> 00:16:59,440 Speaker 5: you can really start layering in for the ones that 365 00:16:59,520 --> 00:17:01,920 Speaker 5: really want to. I've deep into the for the nerds. 366 00:17:01,960 --> 00:17:05,240 Speaker 4: It's for the information rich exactly. 367 00:17:05,320 --> 00:17:07,880 Speaker 5: It's the strategy of tennis. It really should be an 368 00:17:07,880 --> 00:17:09,920 Speaker 5: interesting way to slice and dice a match. 369 00:17:10,160 --> 00:17:11,080 Speaker 4: Huh. 370 00:17:11,160 --> 00:17:14,520 Speaker 3: It's remarkable how the USTA is leveraging AI to enhance 371 00:17:14,560 --> 00:17:19,720 Speaker 3: fan engagement and deliver immersive experiences both on site and online. 372 00:17:20,240 --> 00:17:25,800 Speaker 3: Brian's emphasis on storytelling really underscores the evolution of sports marketing. 373 00:17:26,440 --> 00:17:30,240 Speaker 3: The slam Chakra feature particularly caught my attention. It's essentially 374 00:17:30,280 --> 00:17:33,600 Speaker 3: bringing the excitement of a tennis match to life in 375 00:17:33,640 --> 00:17:37,800 Speaker 3: your palm, moment by moment. As someone who appreciates the 376 00:17:37,880 --> 00:17:41,639 Speaker 3: narrative intricacies of sports, I find it compelling how AI 377 00:17:41,760 --> 00:17:45,600 Speaker 3: helps predict and analyze matches in real time. 378 00:17:46,600 --> 00:17:49,000 Speaker 4: Tell me about the AI commentary feature. 379 00:17:49,200 --> 00:17:52,840 Speaker 5: Yeah, I know. I mentioned AI highlights back in twenty eighteen. 380 00:17:52,920 --> 00:17:55,919 Speaker 5: It's now progressed for us. And again, if we go 381 00:17:56,040 --> 00:17:58,960 Speaker 5: back to before we had a highlights to have a 382 00:17:59,040 --> 00:18:02,199 Speaker 5: highlight ready for this was a video editor cutting the 383 00:18:02,280 --> 00:18:05,920 Speaker 5: highlight and getting voiced over and then being published aside, 384 00:18:05,960 --> 00:18:09,600 Speaker 5: and it took probably an hour plus for that highlight 385 00:18:09,640 --> 00:18:13,320 Speaker 5: to really be created. Now with AI commentary, not only 386 00:18:13,320 --> 00:18:16,440 Speaker 5: are we creating and cutting the highlights using our AI technology, 387 00:18:16,480 --> 00:18:18,679 Speaker 5: but it's now using all the data points that we 388 00:18:18,720 --> 00:18:20,920 Speaker 5: have around the match, whether it's our live scoring data, 389 00:18:21,320 --> 00:18:24,679 Speaker 5: our ball tra directory data, etc. And it's really creating 390 00:18:24,720 --> 00:18:28,199 Speaker 5: a script that helped storytell around that match. That's all 391 00:18:28,280 --> 00:18:32,159 Speaker 5: using Watson X technology and then using text to speech 392 00:18:32,240 --> 00:18:34,720 Speaker 5: we're able to actually then create the commentary on top 393 00:18:34,760 --> 00:18:38,040 Speaker 5: of that, which all happens now within minutes. So our 394 00:18:38,040 --> 00:18:40,960 Speaker 5: team's able to now create fully voiced highlights for every 395 00:18:40,960 --> 00:18:44,159 Speaker 5: men's and women's singles match to our site within minutes. 396 00:18:44,960 --> 00:18:47,479 Speaker 4: So I know there's a new feature you're working on 397 00:18:47,520 --> 00:18:51,760 Speaker 4: for this year called match reports. What are match reports? 398 00:18:52,080 --> 00:18:56,679 Speaker 5: It's our ability to succicktly tell the story of a match, 399 00:18:57,160 --> 00:19:00,680 Speaker 5: so everything happens in five hours within that match down 400 00:19:00,720 --> 00:19:04,680 Speaker 5: to a couple paragraphs that really helps a user understand 401 00:19:04,720 --> 00:19:08,320 Speaker 5: or a fan understand what just happened. Again, some key 402 00:19:08,359 --> 00:19:11,919 Speaker 5: stats what's upcoming really help us with that storytelling. In 403 00:19:11,960 --> 00:19:14,600 Speaker 5: the past, when we have twenty two courts happening at 404 00:19:14,600 --> 00:19:16,800 Speaker 5: a certain time, we would have to pick and choose 405 00:19:16,880 --> 00:19:19,320 Speaker 5: which stories we think or which matches we think are 406 00:19:19,320 --> 00:19:21,359 Speaker 5: going to have the best stories, and that's a really 407 00:19:21,400 --> 00:19:24,320 Speaker 5: hard thing to predict from an editorial perspective. With our 408 00:19:24,359 --> 00:19:26,720 Speaker 5: match reports now we'll be able to have full coverage 409 00:19:26,720 --> 00:19:28,680 Speaker 5: of every single match during the main draw. 410 00:19:29,440 --> 00:19:32,080 Speaker 4: So, of course I want to talk about jeneritive AI. 411 00:19:32,320 --> 00:19:35,199 Speaker 4: How could we not talk about generative Of course, what 412 00:19:35,240 --> 00:19:36,600 Speaker 4: are you working on with jenitive AI? 413 00:19:36,960 --> 00:19:39,320 Speaker 5: So match reports is the prime example of it. So 414 00:19:39,440 --> 00:19:42,520 Speaker 5: Match Reports will be completely using Watson next genera of 415 00:19:42,520 --> 00:19:47,080 Speaker 5: AI technology, And really again to us, it's how can 416 00:19:47,119 --> 00:19:50,399 Speaker 5: we do that storytelling at scale? Tennis is such a 417 00:19:50,480 --> 00:19:53,800 Speaker 5: data rich sport. All sports have data, but tennis has 418 00:19:53,840 --> 00:19:55,960 Speaker 5: a lot of shots and different shot types and ball 419 00:19:56,000 --> 00:20:00,560 Speaker 5: trajectory and live scoring data and umpire chair data and 420 00:20:00,560 --> 00:20:03,920 Speaker 5: and all that. Factoring in jeneral of AI really helps 421 00:20:04,000 --> 00:20:06,879 Speaker 5: us take some of that structured and unstructured data really 422 00:20:07,280 --> 00:20:10,720 Speaker 5: one organize it in a way, but then help us 423 00:20:11,000 --> 00:20:14,119 Speaker 5: quickly tell that story at scale to all of our fans, 424 00:20:14,320 --> 00:20:17,040 Speaker 5: and I think we're really just starting to scratch at 425 00:20:17,040 --> 00:20:20,320 Speaker 5: some of the capabilities, and we're really excited about where 426 00:20:20,320 --> 00:20:22,320 Speaker 5: we're being, but we also see the opportunity of even 427 00:20:22,600 --> 00:20:24,840 Speaker 5: how we can grow to new fans and new fans 428 00:20:24,880 --> 00:20:27,320 Speaker 5: around the world using jener of AI in the future. 429 00:20:28,640 --> 00:20:32,400 Speaker 4: So I'm curious, and you alluded to this a moment ago, 430 00:20:32,400 --> 00:20:34,040 Speaker 4: but I'd like to talk a little bit more about 431 00:20:34,040 --> 00:20:37,760 Speaker 4: it because it seems interesting as a technical problem. Right, 432 00:20:37,960 --> 00:20:43,280 Speaker 4: is the nature of turning tennis matches into stories, which 433 00:20:43,280 --> 00:20:45,679 Speaker 4: is fundamentally what we're talking about here in different ways 434 00:20:45,720 --> 00:20:50,960 Speaker 4: in different media, is about taking both structured data, right 435 00:20:51,080 --> 00:20:54,959 Speaker 4: like the stats who you know, points stats matches, and 436 00:20:55,160 --> 00:20:58,840 Speaker 4: also unstructured data, right like commentary and articles and the 437 00:20:58,920 --> 00:21:03,480 Speaker 4: kind of fuzzier parts of storytelling. And so I'm curious 438 00:21:03,600 --> 00:21:06,800 Speaker 4: how AI kind of helps you manage both the structured 439 00:21:06,840 --> 00:21:07,960 Speaker 4: and the unstructured data. 440 00:21:08,440 --> 00:21:12,280 Speaker 5: Yeah. So, I think the structured data is pretty self experimentatory, 441 00:21:12,600 --> 00:21:14,359 Speaker 5: but when you get into the unstructured data and some 442 00:21:14,440 --> 00:21:16,320 Speaker 5: of the punditry, that's where you get more of the 443 00:21:16,359 --> 00:21:19,760 Speaker 5: opinion pieces into it. Like a specific player matchup, this 444 00:21:19,840 --> 00:21:22,679 Speaker 5: player always plays well against so and so, or as 445 00:21:22,680 --> 00:21:24,520 Speaker 5: they play always played well at night, or they're a 446 00:21:24,520 --> 00:21:28,199 Speaker 5: fan favorite and the crowd, you know, adrenaline and the 447 00:21:28,320 --> 00:21:30,640 Speaker 5: crowd being behind you can really motivate you to play 448 00:21:30,680 --> 00:21:34,000 Speaker 5: a lot better. So it pulls in all those unstructured 449 00:21:34,000 --> 00:21:36,800 Speaker 5: pieces and helps us really put some more rigor around 450 00:21:36,800 --> 00:21:39,439 Speaker 5: it and help add and enrich our storytelling with it. 451 00:21:39,840 --> 00:21:43,760 Speaker 4: And so I'm curious when you're starting to use generative AI, 452 00:21:44,080 --> 00:21:46,840 Speaker 4: you know, over the past few years, like, what were 453 00:21:46,880 --> 00:21:48,239 Speaker 4: your concerns going into that. 454 00:21:48,640 --> 00:21:53,119 Speaker 5: I think our biggest concern is ensuring that one factually 455 00:21:53,400 --> 00:21:55,000 Speaker 5: it is correct, because it's only as good as the 456 00:21:55,040 --> 00:21:57,080 Speaker 5: data you feed in. And how do you really ensure 457 00:21:57,119 --> 00:21:59,760 Speaker 5: that your model's working right and that the output and 458 00:21:59,800 --> 00:22:02,600 Speaker 5: the data you're feeding it matches the output, and how 459 00:22:02,600 --> 00:22:04,719 Speaker 5: do you do that at scale? So we do have 460 00:22:04,760 --> 00:22:07,920 Speaker 5: a lot of human intervention. That's where the IBM consulting team, 461 00:22:08,040 --> 00:22:10,320 Speaker 5: they're on site with us for those full three weeks 462 00:22:10,359 --> 00:22:13,760 Speaker 5: really helping us review everything and we're constantly learning, especially 463 00:22:13,760 --> 00:22:16,480 Speaker 5: early in the tournament. And I would say the other 464 00:22:16,760 --> 00:22:19,399 Speaker 5: big concern, again it goes around to the data, is 465 00:22:19,680 --> 00:22:22,359 Speaker 5: what data do we have available that is trustworthy? So 466 00:22:22,560 --> 00:22:24,560 Speaker 5: you know, we are feel very confident with the data 467 00:22:24,600 --> 00:22:26,200 Speaker 5: that comes off of court, but when we get into 468 00:22:26,240 --> 00:22:29,960 Speaker 5: that unstructured piece, what are the right data sources? How 469 00:22:29,960 --> 00:22:32,320 Speaker 5: do we validate those data sources and how do we 470 00:22:32,760 --> 00:22:35,560 Speaker 5: ensure that they're accurate Because if the data that has 471 00:22:35,600 --> 00:22:37,240 Speaker 5: to go in has to be accurate for the for 472 00:22:37,280 --> 00:22:38,240 Speaker 5: the output. 473 00:22:38,240 --> 00:22:40,639 Speaker 4: So how do you do that? That's the concern? How 474 00:22:40,960 --> 00:22:41,800 Speaker 4: how do you address it? 475 00:22:41,960 --> 00:22:44,439 Speaker 5: Yeah, so I think there's there's a number of tools 476 00:22:44,480 --> 00:22:47,159 Speaker 5: that we use all within the Watson X umbrella. We 477 00:22:47,240 --> 00:22:49,919 Speaker 5: do a lot of training with the IBM team, so 478 00:22:49,960 --> 00:22:53,600 Speaker 5: we have to constantly train and retrain that model. I 479 00:22:53,600 --> 00:22:56,480 Speaker 5: think the other piece that we're doing is again as 480 00:22:56,520 --> 00:22:59,240 Speaker 5: we're creating that content and we have the IBM consulting 481 00:22:59,280 --> 00:23:01,719 Speaker 5: team on site helping us with that, is as we 482 00:23:01,760 --> 00:23:04,840 Speaker 5: see things and we see outputs, it's refeeding that back 483 00:23:04,840 --> 00:23:06,840 Speaker 5: into the model to make it better for the next time. 484 00:23:06,960 --> 00:23:10,480 Speaker 5: So it's a constantly learning process that we're undergoing. 485 00:23:11,040 --> 00:23:14,800 Speaker 4: So I want to talk about scale. Yes, you have 486 00:23:14,960 --> 00:23:18,720 Speaker 4: like what twenty two different courts with matches going all 487 00:23:18,760 --> 00:23:22,440 Speaker 4: at the same time. You're trying to, you know, approximately 488 00:23:22,800 --> 00:23:26,199 Speaker 4: instantly generate summaries of all these matches in something like 489 00:23:26,240 --> 00:23:31,080 Speaker 4: real time, and I'm curious in particular how the IBM 490 00:23:31,160 --> 00:23:35,280 Speaker 4: models you're using, the IBM Granite models are helping you scale. 491 00:23:35,840 --> 00:23:38,560 Speaker 5: Yeah. So I think one of the big learnings we 492 00:23:38,640 --> 00:23:42,320 Speaker 5: had with IBM granted models too is that we're able 493 00:23:42,359 --> 00:23:45,199 Speaker 5: to run it, you know, against last year's tournaments and 494 00:23:45,240 --> 00:23:48,520 Speaker 5: see what the expected outputs could be and really help 495 00:23:48,600 --> 00:23:50,800 Speaker 5: train that model heading into the tournament. Because as we 496 00:23:50,840 --> 00:23:53,240 Speaker 5: talked about in the beginning, is we can plan, plan, 497 00:23:53,240 --> 00:23:55,359 Speaker 5: and plan, but once two players get on court, the 498 00:23:55,400 --> 00:23:58,040 Speaker 5: outcome is unknown. So how do we really run it 499 00:23:58,080 --> 00:24:00,879 Speaker 5: through its paces and really make sure that whatever that 500 00:24:00,920 --> 00:24:03,120 Speaker 5: outcome could be and whatever that scenario is, whether it's 501 00:24:03,520 --> 00:24:06,879 Speaker 5: a fifth set tie break that happens, or maybe there's 502 00:24:06,920 --> 00:24:09,359 Speaker 5: a you know, a fault in the match or something 503 00:24:09,400 --> 00:24:12,720 Speaker 5: that we're not anticipating, that we have that accounted for 504 00:24:12,760 --> 00:24:14,800 Speaker 5: and that the a won't throw off that output. So 505 00:24:14,840 --> 00:24:19,000 Speaker 5: we really try to think through every scenario, which is 506 00:24:19,040 --> 00:24:22,520 Speaker 5: sometimes difficult, right because again live sports is the unknown 507 00:24:22,600 --> 00:24:24,800 Speaker 5: is the unknown that's what makes it fun. We do 508 00:24:24,960 --> 00:24:27,639 Speaker 5: spend a lot of time thinking through potential scenarios and 509 00:24:27,720 --> 00:24:29,919 Speaker 5: ensuring that we have the right data sets and the 510 00:24:30,000 --> 00:24:33,080 Speaker 5: model to predict that tell. 511 00:24:32,880 --> 00:24:36,159 Speaker 4: Me about match reports and the generative AI model you're 512 00:24:36,240 --> 00:24:36,720 Speaker 4: using for that. 513 00:24:37,400 --> 00:24:39,880 Speaker 5: Yeah, so match reports will be new for us this year, 514 00:24:39,920 --> 00:24:42,440 Speaker 5: So we're in testing right now, so we're really excited 515 00:24:42,480 --> 00:24:44,800 Speaker 5: around it. But the model that we'll be able to 516 00:24:44,880 --> 00:24:48,280 Speaker 5: use using Watson X will use a bunch of different 517 00:24:48,320 --> 00:24:51,320 Speaker 5: parts of the suite of tools A meaning that again 518 00:24:51,359 --> 00:24:54,000 Speaker 5: of taking some of that punditry and the unstructured data 519 00:24:54,040 --> 00:24:57,359 Speaker 5: and the editorial spend, take our structured data as well. 520 00:24:57,680 --> 00:25:00,800 Speaker 5: And really what we're working on right now is figuring 521 00:25:00,800 --> 00:25:04,080 Speaker 5: out the right prompts for the AI to really ensure 522 00:25:04,440 --> 00:25:09,080 Speaker 5: that it tells the right structured story, meaning what just happened. Right, 523 00:25:09,240 --> 00:25:12,120 Speaker 5: So our recap is pretty standard. Here's what the data 524 00:25:12,160 --> 00:25:14,440 Speaker 5: is telling us, who won, who lost, how many sets? 525 00:25:14,480 --> 00:25:17,160 Speaker 4: Here's the score the structured data part, that's the easy part. 526 00:25:17,320 --> 00:25:20,080 Speaker 5: Yeah, and then really where it gets exciting is then 527 00:25:20,320 --> 00:25:23,960 Speaker 5: what does this mean? Meaning what's upcoming? So there's all 528 00:25:24,000 --> 00:25:26,240 Speaker 5: these different scenarios when you get into you know, two 529 00:25:26,320 --> 00:25:29,000 Speaker 5: hundred and fifty four players and a large draw. This 530 00:25:29,080 --> 00:25:31,680 Speaker 5: allows us to distill that down and really tell kind 531 00:25:31,680 --> 00:25:34,600 Speaker 5: of what could happen upcoming. The AI helps us do 532 00:25:34,680 --> 00:25:35,560 Speaker 5: that at scale. 533 00:25:35,840 --> 00:25:38,160 Speaker 4: So I want to sort of generalize for a moment 534 00:25:38,200 --> 00:25:41,720 Speaker 4: to talk about kind of you know, broader challenges with 535 00:25:41,760 --> 00:25:44,720 Speaker 4: AI and how you've solved them. You know a lot 536 00:25:44,760 --> 00:25:49,640 Speaker 4: of generative AI pilots fail because the data quality isn't 537 00:25:49,720 --> 00:25:53,000 Speaker 4: high enough, because the risk controls aren't there, and so 538 00:25:53,040 --> 00:25:56,360 Speaker 4: I'm curious how you dealt with those problems and are 539 00:25:56,359 --> 00:25:58,120 Speaker 4: dealing with them data quality. 540 00:25:58,359 --> 00:26:01,080 Speaker 5: Again, we feel calm with the data that is supplied 541 00:26:01,359 --> 00:26:04,359 Speaker 5: from the US open and from the USTA. Right, So 542 00:26:04,440 --> 00:26:07,440 Speaker 5: we have again that's our structure, scoring data and all that. 543 00:26:07,840 --> 00:26:10,240 Speaker 5: I think what we're constantly looking at is when we 544 00:26:10,280 --> 00:26:12,800 Speaker 5: get outside of our known sources and out to third 545 00:26:12,840 --> 00:26:15,159 Speaker 5: parties is that's where a lot of the testing and 546 00:26:15,240 --> 00:26:18,960 Speaker 5: model work happens. So we pull in different data sources 547 00:26:19,080 --> 00:26:22,960 Speaker 5: and really try to work through how it changes that output. Again, 548 00:26:23,080 --> 00:26:24,920 Speaker 5: some of that comes down to where it's an open 549 00:26:24,920 --> 00:26:27,639 Speaker 5: model and the transparency that we have and the learning 550 00:26:27,720 --> 00:26:29,880 Speaker 5: that comes behind it. That's where a lot of that 551 00:26:30,119 --> 00:26:32,520 Speaker 5: confidence can come from, and it comes from a lot 552 00:26:32,560 --> 00:26:36,280 Speaker 5: of testing and feeding it more data. Your second question 553 00:26:36,400 --> 00:26:39,040 Speaker 5: was a little bit more around the output I believe. 554 00:26:38,800 --> 00:26:42,000 Speaker 4: Right, Yeah, and risks right, So risk, I think of 555 00:26:42,080 --> 00:26:44,840 Speaker 4: risk more in terms of output, right, But the obvious 556 00:26:44,880 --> 00:26:47,920 Speaker 4: sphere is like what if it says something wrong? Yeah, 557 00:26:48,000 --> 00:26:51,640 Speaker 4: inflammatory or whatever like that seems scary? 558 00:26:51,880 --> 00:26:54,280 Speaker 5: Yeah, it definitely is, and it's definitely one of our 559 00:26:54,359 --> 00:26:57,080 Speaker 5: largest concerns when we first took this. FORAY, I would 560 00:26:57,080 --> 00:26:59,159 Speaker 5: say a lot of that comes through our work with 561 00:26:59,280 --> 00:27:02,840 Speaker 5: IBM and I consulting team and really ensuring that again 562 00:27:02,880 --> 00:27:05,639 Speaker 5: they're an extension and the partnership there of our team. 563 00:27:06,119 --> 00:27:09,119 Speaker 5: So whenever we are creating let's say it's the match Report, 564 00:27:09,200 --> 00:27:11,520 Speaker 5: and we're going to be creating these excinct articles for 565 00:27:11,600 --> 00:27:15,000 Speaker 5: every single men's and women's single match that happens, is 566 00:27:15,080 --> 00:27:18,080 Speaker 5: all of those will have manual review and people looking 567 00:27:18,160 --> 00:27:20,840 Speaker 5: through them for accuracy to ensure that the model then 568 00:27:20,880 --> 00:27:23,359 Speaker 5: hallucinate or make up a factor or fill in the 569 00:27:23,400 --> 00:27:25,919 Speaker 5: gaps from things like that. That's the first step. And 570 00:27:25,920 --> 00:27:28,560 Speaker 5: then also when our editorial team goes to publish those 571 00:27:28,600 --> 00:27:31,000 Speaker 5: of the website, they're going to be checking it as well. 572 00:27:31,040 --> 00:27:34,080 Speaker 5: So there are manual interventions throughout that to really check 573 00:27:34,119 --> 00:27:37,359 Speaker 5: that model. But we feel that the ability to do 574 00:27:37,400 --> 00:27:39,760 Speaker 5: it at scale and with us more to check that 575 00:27:40,119 --> 00:27:42,440 Speaker 5: is the efficiency problem that we've been looking to solve. 576 00:27:43,240 --> 00:27:46,439 Speaker 4: So the USTA and IBM have been working together on 577 00:27:46,720 --> 00:27:49,439 Speaker 4: digital innovation for like thirty years from you know, the 578 00:27:49,480 --> 00:27:54,200 Speaker 4: first website, yes for the USTA until now. So that's 579 00:27:54,240 --> 00:27:57,960 Speaker 4: the past thirty years. If you look ahead, what's the next. 580 00:27:57,680 --> 00:28:00,440 Speaker 5: Thirty thirty years is a really long time? 581 00:28:00,880 --> 00:28:01,200 Speaker 2: Agree? 582 00:28:01,560 --> 00:28:05,119 Speaker 6: Yeah, I think you know where I get excited, and 583 00:28:05,320 --> 00:28:07,480 Speaker 6: I think I alluded to it in the beginning about 584 00:28:07,480 --> 00:28:09,760 Speaker 6: how I feel like we're just scratching at the surface, 585 00:28:09,840 --> 00:28:12,000 Speaker 6: especially with Journati of Ai, and where I see it 586 00:28:12,080 --> 00:28:15,200 Speaker 6: going is there's a lot of different fans out there, 587 00:28:15,480 --> 00:28:17,280 Speaker 6: and we're also very kindness in the us OP and 588 00:28:17,280 --> 00:28:19,679 Speaker 6: that we're a worldwide event, and that there's a lot 589 00:28:19,720 --> 00:28:23,240 Speaker 6: of different fans that were not necessary creating content for 590 00:28:23,560 --> 00:28:27,359 Speaker 6: bespoke meaning in their native language or maybe it's in 591 00:28:27,359 --> 00:28:29,360 Speaker 6: that native players language and things like that. 592 00:28:29,480 --> 00:28:32,760 Speaker 5: Is where I get excited is we've seen immense growth 593 00:28:32,760 --> 00:28:35,080 Speaker 5: with a Highlights and the ability to now do highlights 594 00:28:35,080 --> 00:28:38,040 Speaker 5: at scale. Is the ability for us to start creating 595 00:28:38,120 --> 00:28:42,320 Speaker 5: content in different languages, maybe covering different parts of the match. 596 00:28:42,360 --> 00:28:44,760 Speaker 5: So maybe you do have that stats junkie you really wants, 597 00:28:44,880 --> 00:28:47,880 Speaker 5: just it's the fastest serve and here's the deep insights 598 00:28:47,960 --> 00:28:50,520 Speaker 5: versus the casual fan who's looking for more of the 599 00:28:50,600 --> 00:28:54,120 Speaker 5: storytelling around how a player trains and what leading up 600 00:28:54,160 --> 00:28:56,040 Speaker 5: to it was like and what it means for them 601 00:28:56,480 --> 00:28:59,320 Speaker 5: afterwards and things like that. A lot of that takes 602 00:28:59,320 --> 00:29:01,440 Speaker 5: a lot of time. Now we're able to solve that 603 00:29:01,480 --> 00:29:04,560 Speaker 5: efficiency problem and do it in multiple languages, we can 604 00:29:04,600 --> 00:29:08,200 Speaker 5: really create I want to say, personalized content to a 605 00:29:08,240 --> 00:29:11,480 Speaker 5: lot more fans all around the world, which again helps 606 00:29:11,560 --> 00:29:14,040 Speaker 5: us grow the sport of tennis great. 607 00:29:14,880 --> 00:29:18,000 Speaker 4: So I want to finish with a speed round. Okay, 608 00:29:18,200 --> 00:29:18,840 Speaker 4: are you ready? 609 00:29:18,960 --> 00:29:19,840 Speaker 5: I am ready? 610 00:29:19,880 --> 00:29:23,400 Speaker 4: Okay, first thing that comes to mind, complete this sentence. 611 00:29:24,040 --> 00:29:26,160 Speaker 4: In five years, AI will. 612 00:29:26,600 --> 00:29:29,120 Speaker 5: Transform many parts of the business. 613 00:29:29,360 --> 00:29:33,600 Speaker 4: What is the number one thing that people misunderstand about AI? 614 00:29:34,160 --> 00:29:38,320 Speaker 5: That it's supplemental, not replacing, meaning that it helps it 615 00:29:38,360 --> 00:29:42,440 Speaker 5: with efficiencies, but it doesn't necessarily replace the creativity. 616 00:29:43,120 --> 00:29:46,800 Speaker 4: Right now, what advice would you give yourself ten years 617 00:29:46,840 --> 00:29:49,680 Speaker 4: ago to better prepare you for today? 618 00:29:50,760 --> 00:29:54,280 Speaker 5: I think it would have been, especially now that we're 619 00:29:54,320 --> 00:29:56,800 Speaker 5: able to take so much of that unstructured data and 620 00:29:57,280 --> 00:29:59,840 Speaker 5: pass content that we were created to help tell st 621 00:30:01,000 --> 00:30:03,840 Speaker 5: was to I want to say archive more of that 622 00:30:03,920 --> 00:30:05,800 Speaker 5: in a way that we could be using that to 623 00:30:05,880 --> 00:30:09,360 Speaker 5: help pull from that now. So you know, we've seen 624 00:30:09,440 --> 00:30:11,480 Speaker 5: kind of a change in the guard from some of 625 00:30:11,520 --> 00:30:14,440 Speaker 5: our star players to now new and up and comers, 626 00:30:14,440 --> 00:30:16,480 Speaker 5: and it would be really fascinating to me if there 627 00:30:16,520 --> 00:30:19,239 Speaker 5: was a way to to cross sections some of that 628 00:30:19,360 --> 00:30:22,120 Speaker 5: and saying like what tra directories are certain up and 629 00:30:22,160 --> 00:30:25,840 Speaker 5: coming players maybe filing from others. So it's more I 630 00:30:25,880 --> 00:30:27,760 Speaker 5: wish we kept more of the content. 631 00:30:27,480 --> 00:30:32,400 Speaker 4: We created back fave the data exactly. Well are you 632 00:30:32,440 --> 00:30:33,360 Speaker 4: saving it all now? 633 00:30:33,640 --> 00:30:35,760 Speaker 5: Oh yeah, one hundred percent learned our lesson? 634 00:30:35,880 --> 00:30:36,640 Speaker 2: Yes, yes. 635 00:30:37,280 --> 00:30:39,640 Speaker 4: So on the business side of AI, what do you 636 00:30:39,680 --> 00:30:40,880 Speaker 4: think is the next big thing? 637 00:30:41,520 --> 00:30:44,800 Speaker 5: I alluded to it earlier. I think it's personalization and 638 00:30:44,840 --> 00:30:48,440 Speaker 5: getting content that's catered to you at scale, whether you 639 00:30:48,480 --> 00:30:51,480 Speaker 5: know that's across the sports sphere or or any type 640 00:30:51,520 --> 00:30:54,720 Speaker 5: of written content or or news content. I feel like 641 00:30:55,480 --> 00:30:58,800 Speaker 5: the ability to really get contentated to the type of 642 00:30:58,880 --> 00:31:01,600 Speaker 5: fan you are and the insight you have is where 643 00:31:01,600 --> 00:31:02,200 Speaker 5: we're all headed. 644 00:31:03,280 --> 00:31:07,120 Speaker 4: And in terms of your non work life, how do 645 00:31:07,160 --> 00:31:08,760 Speaker 4: you use AI day to day? 646 00:31:09,000 --> 00:31:11,320 Speaker 5: It's funny, I was just having this conversation with a 647 00:31:11,360 --> 00:31:14,640 Speaker 5: friend the other day, and we were talking about sometimes 648 00:31:14,640 --> 00:31:17,720 Speaker 5: when you're starting something new, the hardest thing to do 649 00:31:17,840 --> 00:31:20,120 Speaker 5: is you have a blank piece of paper or a thought, 650 00:31:20,160 --> 00:31:24,280 Speaker 5: and how do you get started. Sometimes with these generative models, 651 00:31:24,600 --> 00:31:26,200 Speaker 5: the easiest thing and the best thing you can do 652 00:31:26,320 --> 00:31:28,960 Speaker 5: is it helps you get started. Meaning it may not 653 00:31:29,000 --> 00:31:30,920 Speaker 5: be one hundred percent with that first prompt, but it's 654 00:31:30,960 --> 00:31:34,280 Speaker 5: that efficiency of whether it's an outline for a new idea, 655 00:31:34,480 --> 00:31:36,680 Speaker 5: or it's a marketing brief you have to write, or 656 00:31:36,880 --> 00:31:39,000 Speaker 5: sometimes even if it's an email you have to write 657 00:31:39,160 --> 00:31:41,360 Speaker 5: for a personal something and you're not sure how to 658 00:31:41,400 --> 00:31:43,520 Speaker 5: word it the right way. It allows you to have 659 00:31:43,720 --> 00:31:45,680 Speaker 5: a start and then you can edit from there. So 660 00:31:45,720 --> 00:31:48,720 Speaker 5: again going back to my efficiency point, it helps you 661 00:31:48,800 --> 00:31:49,360 Speaker 5: become more. 662 00:31:49,240 --> 00:31:51,600 Speaker 4: Efficient, solves the blank page problem. 663 00:31:51,800 --> 00:31:52,200 Speaker 5: It does. 664 00:31:53,560 --> 00:31:55,280 Speaker 4: Brian, it was great to talk with you. Thank you 665 00:31:55,280 --> 00:31:56,120 Speaker 4: so much for your time. 666 00:31:56,240 --> 00:31:57,720 Speaker 5: Yeah, this was fun. Thanks for having me. 667 00:32:00,120 --> 00:32:02,400 Speaker 3: Huge thanks to Jacob and Brian for the deep dive 668 00:32:02,720 --> 00:32:06,280 Speaker 3: into the cutting edge innovations transforming the game of tennis. 669 00:32:06,880 --> 00:32:09,800 Speaker 3: Brian shed light on how the US opens partnership with 670 00:32:09,840 --> 00:32:15,120 Speaker 3: IBM is harnessing data driven insights to reshape storytelling in sports, 671 00:32:15,520 --> 00:32:21,040 Speaker 3: from AI generated commentary to Match reports. As we look ahead, 672 00:32:21,120 --> 00:32:25,880 Speaker 3: I'm excited about the possibilities for personalizing content and reaching 673 00:32:25,960 --> 00:32:30,000 Speaker 3: fans in new ways. The future of AI promises more 674 00:32:30,040 --> 00:32:38,360 Speaker 3: than just efficiency. It's about enhancing fan experiences worldwide. Smart 675 00:32:38,360 --> 00:32:42,080 Speaker 3: Talks with IBM is produced by Matt Romano, Joey Fishground, 676 00:32:42,280 --> 00:32:46,320 Speaker 3: and Jacob Goldstein. We're edited by Lydia jen Kott. Our 677 00:32:46,360 --> 00:32:51,040 Speaker 3: engineers are Sarah Bruger and Ben Tolliday. Theme song by Gramascow. 678 00:32:52,000 --> 00:32:54,880 Speaker 3: Special thanks to the eight Bar and IBM teams, as 679 00:32:54,880 --> 00:32:58,400 Speaker 3: well as the Pushkin marketing team. Smart Talks with IBM 680 00:32:58,640 --> 00:33:02,960 Speaker 3: is a production of Pushkin Indie and Ruby Studio at iHeartMedia. 681 00:33:03,480 --> 00:33:07,680 Speaker 3: To find more Pushkin podcasts, listen on the iHeartRadio app, 682 00:33:07,960 --> 00:33:13,720 Speaker 3: Apple Podcasts, or wherever you listen to podcasts. I'm Malcolm Gladwell. 683 00:33:14,040 --> 00:33:17,760 Speaker 3: This is a paid advertisement from IBM. The conversations on 684 00:33:17,800 --> 00:33:36,800 Speaker 3: this podcast don't necessarily represent IBM's positions, strategies, or opinions.