WEBVTT - Smart Talks with IBM: An AI advantage for the US Open

0:00:04.880 --> 0:00:10.840
<v Speaker 1>Welcome to Tech Stuff, a production from iHeartRadio.

0:00:12.560 --> 0:00:14.200
<v Speaker 2>Today, we are witnessed.

0:00:13.720 --> 0:00:16.960
<v Speaker 1>To one of those rare moments in history, the rise

0:00:17.079 --> 0:00:20.880
<v Speaker 1>of an innovative technology with the potential to radically transform

0:00:20.920 --> 0:00:26.640
<v Speaker 1>business and society forever. That technology, of course, is artificial intelligence,

0:00:26.840 --> 0:00:29.600
<v Speaker 1>and it's the central focus for this new season of

0:00:29.640 --> 0:00:33.720
<v Speaker 1>Smart Talks with IBM. Join hosts from your favorite Pushkin

0:00:33.840 --> 0:00:37.360
<v Speaker 1>podcasts as a talk with industry experts and leaders to

0:00:37.440 --> 0:00:41.199
<v Speaker 1>explore how businesses can integrate AI into their workflows and

0:00:41.360 --> 0:00:44.479
<v Speaker 1>help drive real change in this new era of AI,

0:00:45.040 --> 0:00:47.680
<v Speaker 1>and of course, host Malcolm Gladwell will be there to

0:00:47.720 --> 0:00:50.320
<v Speaker 1>guide you through the season and throw in his two

0:00:50.360 --> 0:00:53.400
<v Speaker 1>cents as well. Look out for new episodes of Smart

0:00:53.400 --> 0:00:56.600
<v Speaker 1>Talks with IBM every other week on the iHeartRadio app,

0:00:56.800 --> 0:01:00.400
<v Speaker 1>Apple Podcasts, or wherever you get your podcasts, and learn

0:01:00.440 --> 0:01:13.640
<v Speaker 1>more at IBM dot com, slash smart Talks.

0:01:11.200 --> 0:01:20.720
<v Speaker 3>Pushkin Hello, Hello, Welcome to Smart Talks with IBM, a

0:01:20.800 --> 0:01:25.800
<v Speaker 3>podcast from Pushkin Industries, iHeartRadio and IBM. I'm Malcolm Gladwell.

0:01:26.520 --> 0:01:29.839
<v Speaker 3>This season, we're diving back into the world of artificial intelligence,

0:01:30.240 --> 0:01:33.080
<v Speaker 3>but with a focus on the powerful concept of open

0:01:33.480 --> 0:01:38.480
<v Speaker 3>its possibilities implications and misconceptions. We'll look at openness from

0:01:38.520 --> 0:01:41.440
<v Speaker 3>a variety of angles and explore how the concept is

0:01:41.480 --> 0:01:45.360
<v Speaker 3>already reshaping industries, ways of doing business, and our very

0:01:45.440 --> 0:01:50.680
<v Speaker 3>notion of what's possible. I'm particularly excited for today's guest,

0:01:51.000 --> 0:01:55.160
<v Speaker 3>Brian Ryerson. He's Senior director of Digital Strategy at the

0:01:55.280 --> 0:01:58.760
<v Speaker 3>US Tennis Association, helping to oversee one of the most

0:01:59.000 --> 0:02:02.200
<v Speaker 3>iconic events in the world of sports, the US Open.

0:02:03.200 --> 0:02:06.520
<v Speaker 3>Brian sat down with Pushkin's own Jacob Goldstein, host of

0:02:06.560 --> 0:02:11.000
<v Speaker 3>the podcast What's Your Problem. A veteran business journalist, Jacob

0:02:11.040 --> 0:02:13.920
<v Speaker 3>has reported for The Wall Street Journal, the Miami Herald,

0:02:14.160 --> 0:02:17.799
<v Speaker 3>and was a longtime host of the NPR program Planet Money.

0:02:18.960 --> 0:02:22.320
<v Speaker 3>IBM has been the official technology partner of the US

0:02:22.320 --> 0:02:26.519
<v Speaker 3>Tennis Association for more than thirty years, and the more

0:02:26.560 --> 0:02:31.000
<v Speaker 3>recent evolution into generative AI has enhanced the world class

0:02:31.160 --> 0:02:35.560
<v Speaker 3>digital experiences that help more than fifteen million fans from

0:02:35.600 --> 0:02:39.160
<v Speaker 3>all over the world enjoy the US Open Tennis Championships.

0:02:39.600 --> 0:02:43.560
<v Speaker 3>In this episode, we will explore how generative AI is

0:02:43.639 --> 0:02:48.280
<v Speaker 3>being used to generate match insights, spoken commentary for match highlights,

0:02:48.560 --> 0:02:52.399
<v Speaker 3>and postmatch summaries at scale for fans to enjoy through

0:02:52.400 --> 0:02:55.760
<v Speaker 3>the US Open app and website. We'll explore how these

0:02:55.800 --> 0:03:00.240
<v Speaker 3>AI solutions enable the editorial team to cover more of

0:03:00.280 --> 0:03:04.440
<v Speaker 3>the tournament than ever before, bringing fans even closer to

0:03:04.520 --> 0:03:06.880
<v Speaker 3>the game they love, and we'll learn more about one

0:03:06.880 --> 0:03:11.040
<v Speaker 3>of the engines behind this AI powered content creation, a

0:03:11.160 --> 0:03:15.000
<v Speaker 3>large language model from the IBM Granite family, which is

0:03:15.040 --> 0:03:20.320
<v Speaker 3>trained and maintained using the wantsonex AI and data platform. Okay,

0:03:21.040 --> 0:03:21.760
<v Speaker 3>let's dive in.

0:03:23.000 --> 0:03:24.080
<v Speaker 2>Brian, Welcome to the show.

0:03:24.360 --> 0:03:25.800
<v Speaker 4>Thanks for having me. I'm excited to be here.

0:03:26.200 --> 0:03:27.720
<v Speaker 2>Can you say your name and your job.

0:03:28.080 --> 0:03:30.960
<v Speaker 4>Yeah, I'm a Brian Ryerson. I'm senior director of Digital

0:03:30.960 --> 0:03:32.160
<v Speaker 4>Strategy at the USTA.

0:03:32.520 --> 0:03:34.840
<v Speaker 2>Dumb question, what's the USTA.

0:03:34.639 --> 0:03:36.839
<v Speaker 4>The US Tennis Association.

0:03:37.120 --> 0:03:39.720
<v Speaker 2>And tell me about the USTA, Like, what is it?

0:03:40.120 --> 0:03:42.720
<v Speaker 4>Yeah? So the USTA is the governing body of tennis

0:03:42.720 --> 0:03:46.040
<v Speaker 4>in the US. Or mission is to grow the sport

0:03:46.080 --> 0:03:49.440
<v Speaker 4>of tennis across the US at all levels. Really, I

0:03:49.440 --> 0:03:51.600
<v Speaker 4>would say we're more like a health and wellness company

0:03:51.600 --> 0:03:54.200
<v Speaker 4>where tennis is the means to health and wellness. And

0:03:54.240 --> 0:03:56.720
<v Speaker 4>then the US Open is kind of our tenth pole

0:03:56.800 --> 0:03:59.400
<v Speaker 4>event that happens iver Ear and Flushing Meadows and is

0:03:59.400 --> 0:04:01.480
<v Speaker 4>really our chance. It's the showcase the sport of tennis

0:04:01.480 --> 0:04:03.920
<v Speaker 4>at its highest level to fans all around the world.

0:04:04.040 --> 0:04:06.200
<v Speaker 2>Yeah, I mean the US Open. I assume most people

0:04:06.240 --> 0:04:08.400
<v Speaker 2>know this, but it's Grand Slam. It's one of the

0:04:08.440 --> 0:04:11.520
<v Speaker 2>what four biggest tennis tournaments in the world.

0:04:11.640 --> 0:04:14.840
<v Speaker 4>Yes, yeah, every year, we especially the past couple of years,

0:04:14.880 --> 0:04:18.160
<v Speaker 4>we've seen immense growth and you know, we are very

0:04:18.240 --> 0:04:20.280
<v Speaker 4>hopeful this year and our big goals. Have over a

0:04:20.320 --> 0:04:22.720
<v Speaker 4>million fans on site during the three week window this year,

0:04:22.760 --> 0:04:24.960
<v Speaker 4>so it's an amazing event. I always say it's a

0:04:24.960 --> 0:04:27.840
<v Speaker 4>food and wine festival where tennis is the main attraction,

0:04:28.240 --> 0:04:30.479
<v Speaker 4>and it's a really fun, unique atmosphere.

0:04:31.160 --> 0:04:33.000
<v Speaker 2>How did you get into the tennis business.

0:04:33.160 --> 0:04:35.839
<v Speaker 4>It's a great question. It's not where I thought i'd

0:04:35.920 --> 0:04:38.360
<v Speaker 4>end up for especially being there for fourteen years. So

0:04:38.520 --> 0:04:41.880
<v Speaker 4>I was a marketing and technology major in school, and

0:04:42.600 --> 0:04:45.200
<v Speaker 4>I also played college lacrosse and sports was always a

0:04:45.200 --> 0:04:46.880
<v Speaker 4>big part of my life and always wanted to be

0:04:46.880 --> 0:04:49.320
<v Speaker 4>in the sports and entertainment world. I'm here from the

0:04:49.320 --> 0:04:50.880
<v Speaker 4>New York area. This is where I grew up. So

0:04:50.920 --> 0:04:53.320
<v Speaker 4>I moved back home and had a few friends who

0:04:53.400 --> 0:04:56.000
<v Speaker 4>worked there, and I started out more on the number

0:04:56.120 --> 0:04:58.360
<v Speaker 4>side of things and really digital analytics. It was really

0:04:58.360 --> 0:05:01.440
<v Speaker 4>the start of We're in Facebook and Twitter is just

0:05:01.480 --> 0:05:04.800
<v Speaker 4>starting and digital marketing and all of that. And you know,

0:05:04.839 --> 0:05:07.400
<v Speaker 4>I went to my first US Open not really knowing

0:05:07.440 --> 0:05:10.600
<v Speaker 4>what to expect, and again, I think the atmosphere kind

0:05:10.600 --> 0:05:12.960
<v Speaker 4>of captivated me and hooked me in. And I've been

0:05:12.960 --> 0:05:14.080
<v Speaker 4>there now fourteen years.

0:05:14.520 --> 0:05:18.760
<v Speaker 2>And so your title is Digital Director. What does that mean?

0:05:18.800 --> 0:05:19.400
<v Speaker 2>What's your job?

0:05:19.640 --> 0:05:22.640
<v Speaker 4>Yeah, so it's an interesting one because it's tough to

0:05:22.720 --> 0:05:25.240
<v Speaker 4>explain to folks who are not in the weeds on

0:05:25.320 --> 0:05:27.719
<v Speaker 4>all things us Open or even in the sports world.

0:05:27.720 --> 0:05:31.440
<v Speaker 4>But really I oversee all of our consumer facing digital property.

0:05:31.560 --> 0:05:34.040
<v Speaker 4>So that's the us open dot org, our website built

0:05:34.040 --> 0:05:36.920
<v Speaker 4>by IBM, as well as our mobile app. I oversee

0:05:36.920 --> 0:05:42.000
<v Speaker 4>our content strategy, our sponsorship integrations. So really anything consumer

0:05:42.080 --> 0:05:45.120
<v Speaker 4>facing that happens on the web is under my purview.

0:05:45.400 --> 0:05:48.200
<v Speaker 4>Even some of our new platform extensions and gaming and

0:05:48.240 --> 0:05:51.440
<v Speaker 4>things like that. Anything that you can physically interact with

0:05:51.839 --> 0:05:53.240
<v Speaker 4>is kind of under my purview.

0:05:54.000 --> 0:05:57.840
<v Speaker 2>And so you've been there now for fourteen ISSU years,

0:05:58.000 --> 0:06:01.640
<v Speaker 2>which in the digital world is a long time. How

0:06:01.680 --> 0:06:06.040
<v Speaker 2>has that sort of digital experience of sports changed over

0:06:06.080 --> 0:06:06.560
<v Speaker 2>that time.

0:06:06.800 --> 0:06:10.360
<v Speaker 4>Yeah, it's obviously grown digital now, is what we say

0:06:10.360 --> 0:06:12.479
<v Speaker 4>and what my team says. It's the number one way

0:06:12.520 --> 0:06:15.159
<v Speaker 4>to engage with fans that can't make it to the event,

0:06:15.360 --> 0:06:16.880
<v Speaker 4>as well as those fans who are at the event,

0:06:16.920 --> 0:06:18.400
<v Speaker 4>and how do you enrich their stay. So it's really

0:06:18.520 --> 0:06:22.840
<v Speaker 4>kind of you're tackling multiple fan personas. It's the international

0:06:22.880 --> 0:06:25.640
<v Speaker 4>fan who's staying up late to watch in other countries,

0:06:25.680 --> 0:06:28.400
<v Speaker 4>to the fan here who's maybe watching on broadcasts, and

0:06:28.400 --> 0:06:30.640
<v Speaker 4>we go on in a company and enrich that broadcast of

0:06:30.760 --> 0:06:33.080
<v Speaker 4>new stats and insights to the on site fan who

0:06:33.080 --> 0:06:35.320
<v Speaker 4>bought a ticket and maybe doesn't know what match is

0:06:35.320 --> 0:06:38.120
<v Speaker 4>happening on what court. We do have twenty plus courts

0:06:38.120 --> 0:06:40.760
<v Speaker 4>happening at a time with all different matches, So really

0:06:40.760 --> 0:06:43.800
<v Speaker 4>try to help all fans navigate the US Open the

0:06:43.800 --> 0:06:44.640
<v Speaker 4>best way possible.

0:06:45.040 --> 0:06:47.159
<v Speaker 2>And so, like, what are some of the sort of

0:06:47.200 --> 0:06:49.159
<v Speaker 2>problems you're trying to solve. What are some of the

0:06:49.160 --> 0:06:50.480
<v Speaker 2>hard things about your job?

0:06:50.800 --> 0:06:54.920
<v Speaker 4>Yeah, obviously technology changes at a rapid pace, right, So

0:06:55.000 --> 0:06:56.880
<v Speaker 4>I think part of it is how do we stay

0:06:56.880 --> 0:06:59.040
<v Speaker 4>on the forefront of that, and how do we do

0:06:59.080 --> 0:07:01.200
<v Speaker 4>that in the best way and make the best fan

0:07:01.320 --> 0:07:04.800
<v Speaker 4>experiences possible and the best user experience as possible. That's

0:07:04.839 --> 0:07:08.159
<v Speaker 4>always kind of driving factor number one. Then number two,

0:07:08.200 --> 0:07:11.800
<v Speaker 4>it's understanding and listening to our fans and what kind

0:07:11.840 --> 0:07:14.040
<v Speaker 4>of content they want. You hear me talk a lot

0:07:14.040 --> 0:07:16.560
<v Speaker 4>about storytelling. I feel like there's a lot of storytelling

0:07:16.560 --> 0:07:18.480
<v Speaker 4>that happens around the years open that we really want

0:07:18.520 --> 0:07:21.520
<v Speaker 4>to bring to fans. And that can be as simple

0:07:21.600 --> 0:07:24.840
<v Speaker 4>as storytelling of what's happening today and what you should

0:07:24.920 --> 0:07:28.240
<v Speaker 4>be watching too. Maybe it's your favorite players and what's

0:07:28.280 --> 0:07:31.840
<v Speaker 4>going on behind the scenes with them, to even introducing

0:07:32.040 --> 0:07:34.840
<v Speaker 4>I want to say, the casual fans to who they

0:07:34.840 --> 0:07:37.280
<v Speaker 4>should be watching, why they should follow certain players, and

0:07:37.280 --> 0:07:39.320
<v Speaker 4>more bringing that player's story to life.

0:07:39.960 --> 0:07:42.360
<v Speaker 2>Yeah, I mean, I feel like almost the whole point

0:07:42.400 --> 0:07:45.760
<v Speaker 2>of sports is to create stories for us to follow, right,

0:07:45.840 --> 0:07:49.440
<v Speaker 2>Like they're engineered to be stories. It's exactly this thing

0:07:49.520 --> 0:07:51.640
<v Speaker 2>is happening in front of you and there are two

0:07:52.200 --> 0:07:55.160
<v Speaker 2>antagonists and the stakes are high, and you don't know

0:07:55.200 --> 0:07:57.720
<v Speaker 2>how it's gonna end. Like it's built to be a story.

0:07:57.960 --> 0:08:00.240
<v Speaker 4>Yeah, And that's the main challenge of the job is

0:08:00.280 --> 0:08:03.000
<v Speaker 4>you can plan, plan, plan, but once you get on

0:08:03.160 --> 0:08:05.080
<v Speaker 4>two players on court and you don't know what that

0:08:05.120 --> 0:08:07.240
<v Speaker 4>outcome is going to be. It's now sitting and waiting

0:08:07.240 --> 0:08:09.680
<v Speaker 4>and watching and you become a fan yourself. And then

0:08:09.720 --> 0:08:13.120
<v Speaker 4>it's how do you really captivate that story and how

0:08:13.120 --> 0:08:15.600
<v Speaker 4>do you narrate it? How do you like translate up

0:08:15.640 --> 0:08:16.400
<v Speaker 4>to fans.

0:08:16.280 --> 0:08:17.720
<v Speaker 2>And it's like you kind of have to do it

0:08:17.760 --> 0:08:19.760
<v Speaker 2>in real time, right, Like the whole point of sports

0:08:19.840 --> 0:08:20.680
<v Speaker 2>is you don't know what's going.

0:08:20.680 --> 0:08:23.280
<v Speaker 4>To happen exactly, and that's the excitement. And it's also

0:08:23.440 --> 0:08:25.520
<v Speaker 4>there's so many different types of fans. You know, there's

0:08:25.560 --> 0:08:28.080
<v Speaker 4>the fans who want a lot of enriched data and

0:08:28.120 --> 0:08:30.360
<v Speaker 4>their tennis nerds for lack of better of saying it,

0:08:30.400 --> 0:08:32.760
<v Speaker 4>and that they really want to dive deep into the

0:08:32.800 --> 0:08:35.800
<v Speaker 4>intricacies of the game, versus the casual fan who maybe

0:08:35.840 --> 0:08:37.760
<v Speaker 4>just wants more of this high level storyline of what

0:08:37.800 --> 0:08:41.120
<v Speaker 4>does this mean? Why is it important? So it's really

0:08:41.120 --> 0:08:43.680
<v Speaker 4>trying to figure out how to deliver that at scale

0:08:43.679 --> 0:08:46.480
<v Speaker 4>and really help fans get what they're looking for and

0:08:46.520 --> 0:08:47.880
<v Speaker 4>the type of content they're looking for.

0:08:48.200 --> 0:08:51.400
<v Speaker 2>So, are there specific examples of, you know, how fan

0:08:51.520 --> 0:08:56.040
<v Speaker 2>feedback has led to specific features digital features you build?

0:08:56.400 --> 0:08:59.400
<v Speaker 2>Are there like particularly popular features you've come up with,

0:08:59.440 --> 0:09:01.200
<v Speaker 2>Like what are some specifics.

0:09:00.720 --> 0:09:03.240
<v Speaker 4>Yeah, some low hanging fruit type things that came from

0:09:03.280 --> 0:09:06.640
<v Speaker 4>fan feedback. Is simple things sometimes like managing time zones

0:09:06.760 --> 0:09:07.959
<v Speaker 4>when matches start.

0:09:07.960 --> 0:09:11.560
<v Speaker 2>A persistent problem where those of us work across.

0:09:11.280 --> 0:09:15.120
<v Speaker 4>Times exactly, and we do have, like I mentioned, twenty

0:09:15.120 --> 0:09:17.320
<v Speaker 4>plus courts happening at a time, So it's a lot

0:09:17.360 --> 0:09:19.720
<v Speaker 4>to follow, and how do you translate that to a fan,

0:09:20.160 --> 0:09:22.719
<v Speaker 4>whether it's to their native language or to their time

0:09:22.800 --> 0:09:25.080
<v Speaker 4>zone or things like that. So that's one thing that

0:09:25.120 --> 0:09:27.960
<v Speaker 4>came through fan feedback, and another one a three to

0:09:28.040 --> 0:09:30.760
<v Speaker 4>five hour match, especially when you're having twenty plus of

0:09:30.800 --> 0:09:32.880
<v Speaker 4>them happening at a time, is there's too much for

0:09:32.960 --> 0:09:36.520
<v Speaker 4>one person to follow. So how do you start from

0:09:36.520 --> 0:09:40.000
<v Speaker 4>an editorial perspective really helping with that storytelling and guiding

0:09:40.000 --> 0:09:42.400
<v Speaker 4>a fan to like, all right, whether there's an upset

0:09:42.440 --> 0:09:46.040
<v Speaker 4>about to happen, or here's your matches to watch, or

0:09:46.400 --> 0:09:48.280
<v Speaker 4>even some of the predictions we're starting to put in

0:09:48.360 --> 0:09:50.480
<v Speaker 4>is we really want to guide the fan before match,

0:09:50.840 --> 0:09:53.360
<v Speaker 4>here's where you should tune in to even after a

0:09:53.400 --> 0:09:56.640
<v Speaker 4>match of here's what's happened, here's what's important, and we're

0:09:56.679 --> 0:09:58.240
<v Speaker 4>really excited with some of the features we've built in

0:09:58.280 --> 0:10:00.920
<v Speaker 4>the last few years that I would say really helps

0:10:00.960 --> 0:10:02.600
<v Speaker 4>us do that at more scale than what we were

0:10:02.640 --> 0:10:04.960
<v Speaker 4>able to do with just writers following a match and

0:10:05.000 --> 0:10:06.240
<v Speaker 4>covering every single match.

0:10:06.400 --> 0:10:08.679
<v Speaker 2>Uh huh. So I want to talk a little bit

0:10:08.760 --> 0:10:14.560
<v Speaker 2>about the partnership between IBM and the USTA, Like, just

0:10:14.640 --> 0:10:16.360
<v Speaker 2>tell me about the work you do together.

0:10:16.760 --> 0:10:20.000
<v Speaker 4>So, IBM is our official digital and technology partner and

0:10:20.080 --> 0:10:23.079
<v Speaker 4>innovation partner of the US Open They predate me. It's

0:10:23.080 --> 0:10:26.000
<v Speaker 4>a thirty year partnership and it truly is a partnership.

0:10:26.040 --> 0:10:29.040
<v Speaker 4>So I view the IBM consulting team as an extension

0:10:29.120 --> 0:10:32.600
<v Speaker 4>of my USTA team, So we work with them year round.

0:10:32.880 --> 0:10:37.199
<v Speaker 4>They design, develop and deliver the digital properties. They help

0:10:37.280 --> 0:10:40.000
<v Speaker 4>us provide the tools to create content, to do things

0:10:40.000 --> 0:10:42.440
<v Speaker 4>at scale. They help us from stats and information and

0:10:42.760 --> 0:10:45.400
<v Speaker 4>really help us push from an innovation standpoint to make

0:10:45.400 --> 0:10:47.720
<v Speaker 4>sure that we are staying on that cutting edge of technology.

0:10:47.760 --> 0:10:52.280
<v Speaker 4>So I would truly say it's much more than a sponsorship,

0:10:52.280 --> 0:10:55.000
<v Speaker 4>where it's truly a partnership to deliver that fan experience.

0:10:55.440 --> 0:10:58.079
<v Speaker 2>And So, what are some of the specific things that

0:10:58.360 --> 0:10:59.480
<v Speaker 2>you have done with IBM.

0:10:59.720 --> 0:11:03.960
<v Speaker 4>Yeah, so I mean, there's countless ones to talk through. Obviously,

0:11:04.000 --> 0:11:06.600
<v Speaker 4>they thirty years ago. They helped us build our first

0:11:06.600 --> 0:11:10.120
<v Speaker 4>website and it's kind of grown from there over the

0:11:10.240 --> 0:11:12.080
<v Speaker 4>past few years. I would say I think it was

0:11:12.120 --> 0:11:14.680
<v Speaker 4>twenty eighteen as we started AI Highlights, So that was

0:11:14.720 --> 0:11:18.400
<v Speaker 4>really when we were able to have all twenty matches

0:11:18.440 --> 0:11:20.800
<v Speaker 4>going at a single time. We were able to quickly

0:11:20.880 --> 0:11:24.520
<v Speaker 4>deliver succinct highlights to fans to our digital platform, so

0:11:24.520 --> 0:11:26.760
<v Speaker 4>they could see highlights for every single core.

0:11:27.440 --> 0:11:30.360
<v Speaker 2>Is that video highlights? Is that tech summaries? What does

0:11:30.400 --> 0:11:30.840
<v Speaker 2>that mean?

0:11:31.360 --> 0:11:33.840
<v Speaker 4>At the time, it was video highlights, Okay, so it

0:11:33.880 --> 0:11:36.760
<v Speaker 4>was really taking that three to five hour match, let's say,

0:11:36.760 --> 0:11:39.280
<v Speaker 4>and cut it down to a three minute highlight that

0:11:39.400 --> 0:11:41.800
<v Speaker 4>could show up within moments after a match, ending to

0:11:41.840 --> 0:11:44.320
<v Speaker 4>our website and our mobile app, so fans could see

0:11:44.320 --> 0:11:46.000
<v Speaker 4>that all around the world and really kind of get

0:11:46.000 --> 0:11:48.760
<v Speaker 4>that three minute overview what happened in a match?

0:11:48.960 --> 0:11:51.800
<v Speaker 2>And was that AI enabled? Was AI a piece of

0:11:52.240 --> 0:11:52.920
<v Speaker 2>how to do that?

0:11:53.040 --> 0:11:55.680
<v Speaker 4>It was? It was probably our first FORAY into AI.

0:11:56.280 --> 0:12:01.200
<v Speaker 2>Back twenty eighteen is relatively early, yeah, exactly, cly for

0:12:01.320 --> 0:12:03.079
<v Speaker 2>tennis exactly.

0:12:03.200 --> 0:12:06.160
<v Speaker 4>Yeah. It really I want to say opened up our

0:12:06.200 --> 0:12:10.160
<v Speaker 4>ability to one against storyteller but attract new fans too.

0:12:10.240 --> 0:12:12.480
<v Speaker 4>Is video has actually been our number one growth area

0:12:12.600 --> 0:12:14.720
<v Speaker 4>since twenty eighteen, and I think a lot of that

0:12:14.800 --> 0:12:16.480
<v Speaker 4>has to do with the scale of how we deliver

0:12:16.559 --> 0:12:17.440
<v Speaker 4>that content.

0:12:17.440 --> 0:12:20.360
<v Speaker 2>Using AI and being able to deliver the sort of

0:12:20.480 --> 0:12:23.679
<v Speaker 2>video highlight reels at scale.

0:12:23.440 --> 0:12:26.040
<v Speaker 4>Yeah, and do it quickly. Right. We've always had highlights,

0:12:26.040 --> 0:12:28.040
<v Speaker 4>but it was a manual process where you had a

0:12:28.559 --> 0:12:32.280
<v Speaker 4>video ad or cutting through a three hour match, selecting

0:12:32.320 --> 0:12:34.360
<v Speaker 4>the right scene, stitching together it would have to get

0:12:34.400 --> 0:12:37.600
<v Speaker 4>voiced over, et cetera. We really have used AI to

0:12:37.600 --> 0:12:39.320
<v Speaker 4>make it, i want to say, much more efficient and

0:12:39.360 --> 0:12:42.320
<v Speaker 4>speed up that process and deliver it more quickly to

0:12:42.360 --> 0:12:42.840
<v Speaker 4>our fans.

0:12:43.240 --> 0:12:45.160
<v Speaker 2>I mean, it would be a bummer to get scooped

0:12:45.280 --> 0:12:48.400
<v Speaker 2>by whatever NBC News or Yes Pen or whatever. I'm

0:12:48.400 --> 0:12:49.920
<v Speaker 2>sure there are all your partners and you love them.

0:12:49.960 --> 0:12:52.679
<v Speaker 2>Most likely obviously you want to have the video first, right,

0:12:52.760 --> 0:12:53.480
<v Speaker 2>it's your match.

0:12:53.679 --> 0:12:56.160
<v Speaker 4>Yeah, And I think it's also important to us as

0:12:56.280 --> 0:13:00.400
<v Speaker 4>being the USTA is ensuring that it's not just you know,

0:13:00.600 --> 0:13:04.160
<v Speaker 4>the main marquee players, that every player and all those

0:13:04.200 --> 0:13:07.680
<v Speaker 4>storylines and that whether it's you know, the main singles

0:13:07.760 --> 0:13:10.360
<v Speaker 4>draw to our mixed doubles, et cetera. They all need

0:13:10.440 --> 0:13:12.880
<v Speaker 4>highlights and they all have their own stories to tell,

0:13:12.880 --> 0:13:14.640
<v Speaker 4>and how do we do that at scale? It was

0:13:14.679 --> 0:13:17.120
<v Speaker 4>something that before we had that product was not something

0:13:17.120 --> 0:13:17.880
<v Speaker 4>you were able to do.

0:13:18.320 --> 0:13:21.760
<v Speaker 2>Great, So let's let's talk in some more detail about

0:13:21.760 --> 0:13:25.160
<v Speaker 2>what you're working on. Let's start with the app. Tell

0:13:25.200 --> 0:13:28.240
<v Speaker 2>me about the us Open app and the Companion website.

0:13:28.320 --> 0:13:30.560
<v Speaker 4>Yeah, so, so I'll start with the app, and I

0:13:30.600 --> 0:13:33.920
<v Speaker 4>feel like they serve similar needs, but they're a little

0:13:33.960 --> 0:13:37.240
<v Speaker 4>different in their own respective manners. Is the app. Everybody

0:13:37.280 --> 0:13:39.040
<v Speaker 4>has a phone in their hands at this point. The

0:13:39.080 --> 0:13:41.480
<v Speaker 4>app is kind of their guide to when I say

0:13:41.520 --> 0:13:44.120
<v Speaker 4>a million fans on site, we view the app as

0:13:44.200 --> 0:13:46.240
<v Speaker 4>we want that to be their on site guide and

0:13:46.280 --> 0:13:47.720
<v Speaker 4>Companion a million.

0:13:47.840 --> 0:13:50.560
<v Speaker 2>Let's just pause on a million fans on site, right,

0:13:50.600 --> 0:13:54.320
<v Speaker 2>because like a big professional whatever, an NFL game or

0:13:54.360 --> 0:13:57.840
<v Speaker 2>something that's like one hundred thousand, this is ten x that.

0:13:58.200 --> 0:14:01.679
<v Speaker 4>Yeah, and a three week window in a very succinct, tight,

0:14:01.760 --> 0:14:05.800
<v Speaker 4>action packed window. There's a lot of action logistics.

0:14:05.840 --> 0:14:07.560
<v Speaker 2>Okay, so keep going.

0:14:07.679 --> 0:14:10.480
<v Speaker 4>So the app, you know, whether it's finding the schedules,

0:14:10.559 --> 0:14:13.440
<v Speaker 4>the live scores, what's happening on court. That's really the

0:14:13.480 --> 0:14:16.560
<v Speaker 4>focus point of the app, and what we're really focused

0:14:16.600 --> 0:14:18.840
<v Speaker 4>on this year is how do we build in some

0:14:18.880 --> 0:14:21.560
<v Speaker 4>of those match summaries into the app, into our slam

0:14:21.600 --> 0:14:24.880
<v Speaker 4>Tracker experience. So again, before match, that kind of match

0:14:24.920 --> 0:14:27.200
<v Speaker 4>preview of here's maybe if you have a ticket, here's

0:14:27.200 --> 0:14:30.040
<v Speaker 4>what to expect, here's you know are likely to win,

0:14:30.120 --> 0:14:32.600
<v Speaker 4>who we are predicting, so you can kind of get

0:14:32.640 --> 0:14:36.000
<v Speaker 4>some information heading in, and then after the match it's

0:14:36.040 --> 0:14:39.240
<v Speaker 4>more of what just happened, what it means for the

0:14:39.480 --> 0:14:42.240
<v Speaker 4>rest of the draw, who they're playing next, is this

0:14:42.360 --> 0:14:44.520
<v Speaker 4>the first time this has happened, et cetera, and really

0:14:44.600 --> 0:14:47.680
<v Speaker 4>enriching that experience as well. So the app is one

0:14:47.720 --> 0:14:50.080
<v Speaker 4>your guide to what you should be watching, but also

0:14:50.120 --> 0:14:52.840
<v Speaker 4>then giving that insights and context to what's happening on

0:14:52.880 --> 0:14:53.720
<v Speaker 4>that court as you're.

0:14:53.560 --> 0:14:57.200
<v Speaker 2>Watching, like the commentator in your pocket exactly. So you

0:14:57.320 --> 0:14:59.880
<v Speaker 2>used a phrase in there as if I already knew,

0:15:01.000 --> 0:15:02.560
<v Speaker 2>and I love the phrase, but I want you to

0:15:02.600 --> 0:15:05.480
<v Speaker 2>talk more about it. That phrase is slam Tracker.

0:15:05.680 --> 0:15:11.400
<v Speaker 4>Yes, so slam Tracker is our long standing live scores.

0:15:11.440 --> 0:15:13.720
<v Speaker 4>I want to say match center. It is okay, where

0:15:13.960 --> 0:15:16.680
<v Speaker 4>every single data point for every single match lives and

0:15:16.720 --> 0:15:19.960
<v Speaker 4>it really it helps showcase what's happening to match. I say,

0:15:20.000 --> 0:15:22.800
<v Speaker 4>it's our broadcast companion, So if you're watching live, it's

0:15:22.800 --> 0:15:25.360
<v Speaker 4>our in stadium companion. And it's also the best thing

0:15:25.400 --> 0:15:27.440
<v Speaker 4>to have if you aren't able to watch.

0:15:27.560 --> 0:15:29.320
<v Speaker 2>And so like, I'm on the app and there's a

0:15:29.360 --> 0:15:32.240
<v Speaker 2>thing called slam Tracker, and I like, tap slam Tracker.

0:15:32.240 --> 0:15:34.000
<v Speaker 2>What do I see on my phone when I tap

0:15:34.080 --> 0:15:37.400
<v Speaker 2>slam Tracker? You know, midday when the tournament's happening.

0:15:37.480 --> 0:15:39.320
<v Speaker 4>So before match, that's where you get a lot of

0:15:39.360 --> 0:15:41.560
<v Speaker 4>pre match content. That's where those live kind of our

0:15:41.560 --> 0:15:44.920
<v Speaker 4>predictions are. Likelihood to win lives within that. So likelihood

0:15:44.960 --> 0:15:47.560
<v Speaker 4>to win essentially pulls in a bunch of data points.

0:15:47.560 --> 0:15:50.680
<v Speaker 4>So pass matches, how many times these players have played

0:15:50.680 --> 0:15:53.360
<v Speaker 4>each other against each other, even some punditry and other

0:15:53.400 --> 0:15:56.200
<v Speaker 4>written articles that maybe our editorial team put out and

0:15:56.240 --> 0:15:58.880
<v Speaker 4>really kind of puts a prediction out there.

0:15:58.760 --> 0:16:00.800
<v Speaker 2>And so it's just a percentage chance.

0:16:00.800 --> 0:16:03.880
<v Speaker 4>Yes exactly, but it uses millions of data points to

0:16:03.920 --> 0:16:06.200
<v Speaker 4>come up with that. Yes, so it really helps you

0:16:06.280 --> 0:16:09.200
<v Speaker 4>kind of understand what you're getting into for that match.

0:16:09.600 --> 0:16:12.320
<v Speaker 4>During a live match, it is every single point so

0:16:12.720 --> 0:16:15.880
<v Speaker 4>point by point scoring as well as in depth analysis

0:16:15.880 --> 0:16:18.800
<v Speaker 4>in point commentary where also this year have a live

0:16:18.880 --> 0:16:21.840
<v Speaker 4>visualization that accompanies that that will really help bring the

0:16:21.960 --> 0:16:24.360
<v Speaker 4>match together. And what I mean by that is it

0:16:24.480 --> 0:16:27.720
<v Speaker 4>uses our ball tracking technology to really showcase the match

0:16:28.040 --> 0:16:30.960
<v Speaker 4>in near real time, so within seconds delay of where

0:16:30.960 --> 0:16:33.480
<v Speaker 4>the ball is being hit, where the players are, and

0:16:33.520 --> 0:16:36.320
<v Speaker 4>really bring a visualization to life and layered stats and

0:16:36.400 --> 0:16:37.080
<v Speaker 4>data on top of it.

0:16:37.240 --> 0:16:39.120
<v Speaker 2>Uh. Is that sort of like when I'm watching a

0:16:39.200 --> 0:16:41.920
<v Speaker 2>match on TV and there's like a close call as

0:16:41.960 --> 0:16:43.600
<v Speaker 2>the ball in or out and they do that thing

0:16:43.640 --> 0:16:45.640
<v Speaker 2>where they kind of show a sort of video game

0:16:45.720 --> 0:16:47.840
<v Speaker 2>version of where the ball landed. Does it look like that?

0:16:48.120 --> 0:16:50.840
<v Speaker 4>It's like that before every single shot, So it's not

0:16:50.960 --> 0:16:53.440
<v Speaker 4>just those close ones. It's our first foray to bring

0:16:53.480 --> 0:16:54.440
<v Speaker 4>that match to life.

0:16:54.960 --> 0:16:56.840
<v Speaker 2>Huh. And so what do I see on that kind

0:16:56.840 --> 0:16:59.040
<v Speaker 2>of view that I don't see from whatever watching the video?

0:16:59.120 --> 0:17:02.440
<v Speaker 4>Yeah, be able just to see more of the ball

0:17:02.480 --> 0:17:05.239
<v Speaker 4>trajectory and where the ball is being hit, But then

0:17:05.280 --> 0:17:07.840
<v Speaker 4>you can also start layering things and stats and insights

0:17:07.880 --> 0:17:10.560
<v Speaker 4>on top of that. So how many times has player

0:17:10.600 --> 0:17:13.600
<v Speaker 4>A hit the ball on a certain baseline, how fast

0:17:13.600 --> 0:17:16.199
<v Speaker 4>are they hitting it, maybe their serve percentage and a

0:17:16.240 --> 0:17:18.120
<v Speaker 4>certain side of the court, et cetera. So you can

0:17:18.119 --> 0:17:20.600
<v Speaker 4>really start layering in for the ones that really want

0:17:20.640 --> 0:17:21.640
<v Speaker 4>to dive deep into the.

0:17:21.880 --> 0:17:26.000
<v Speaker 2>For the nerds, it's for the it's the information rich exactly.

0:17:26.040 --> 0:17:28.560
<v Speaker 4>It's the strategy of tennis. It really should be an

0:17:28.600 --> 0:17:30.639
<v Speaker 4>interesting way to slice and dice a match.

0:17:30.840 --> 0:17:31.800
<v Speaker 2>Huh.

0:17:31.880 --> 0:17:35.200
<v Speaker 3>It's remarkable how the USTA is leveraging AI to enhance

0:17:35.280 --> 0:17:40.560
<v Speaker 3>fan engagement and deliver immersive experiences both on site and online.

0:17:40.960 --> 0:17:46.480
<v Speaker 3>Brian's emphasis on storytelling really underscores the evolution of sports marketing.

0:17:47.119 --> 0:17:50.920
<v Speaker 3>The slam Choker feature particularly caught my attention. It's essentially

0:17:50.960 --> 0:17:54.280
<v Speaker 3>bringing the excitement of a tennis match to life in

0:17:54.320 --> 0:17:58.520
<v Speaker 3>your palm, moment by moment. As someone who appreciates the

0:17:58.600 --> 0:18:02.320
<v Speaker 3>narrative intricacies of spoil, I find it compelling how AI

0:18:02.440 --> 0:18:06.280
<v Speaker 3>helps predict and analyze matches in real time.

0:18:07.320 --> 0:18:09.719
<v Speaker 2>Tell me about the AI commentary feature.

0:18:09.880 --> 0:18:13.520
<v Speaker 4>Yeah, I know. I mentioned AI highlights back in twenty eighteen.

0:18:13.600 --> 0:18:16.600
<v Speaker 4>It's now progressed for us. And again, if we go

0:18:16.720 --> 0:18:19.679
<v Speaker 4>back to before we had AA highlights, to have a

0:18:19.720 --> 0:18:22.840
<v Speaker 4>highlight ready for the site was a video editor cutting

0:18:22.840 --> 0:18:26.639
<v Speaker 4>the highlight and getting voiced over and then being published aside,

0:18:26.680 --> 0:18:30.280
<v Speaker 4>and it took probably an hour plus for that highlight

0:18:30.320 --> 0:18:34.000
<v Speaker 4>to really be created. Now with AI commentary, not only

0:18:34.040 --> 0:18:37.119
<v Speaker 4>are we creating and cutting the highlights using our AI technology,

0:18:37.160 --> 0:18:39.399
<v Speaker 4>but it's now using all the data points that we

0:18:39.400 --> 0:18:41.680
<v Speaker 4>have around the match, whether it's our live scoring data,

0:18:42.000 --> 0:18:45.399
<v Speaker 4>our ball tro directory data, etc. And it's really creating

0:18:45.400 --> 0:18:48.919
<v Speaker 4>a script that helped storytell around that match. That's all

0:18:49.000 --> 0:18:52.879
<v Speaker 4>using Watson X technology and then using text to speech

0:18:52.920 --> 0:18:55.440
<v Speaker 4>we're able to actually then create the commentary on top

0:18:55.480 --> 0:18:58.719
<v Speaker 4>of that, which all happens now within minutes. So our

0:18:58.760 --> 0:19:01.640
<v Speaker 4>team's able to now create fully voiced highlights for every

0:19:01.680 --> 0:19:04.879
<v Speaker 4>men's and women's singles match to our site within minutes.

0:19:05.640 --> 0:19:08.159
<v Speaker 2>So I know there's a new feature you're working on

0:19:08.200 --> 0:19:12.639
<v Speaker 2>for this year called match reports. Yes, what are match reports?

0:19:12.800 --> 0:19:17.359
<v Speaker 4>It's our ability to succicktly tell the story of a match,

0:19:17.880 --> 0:19:21.359
<v Speaker 4>so everything happens in five hours within that match down

0:19:21.440 --> 0:19:25.320
<v Speaker 4>to a couple paragraphs. That really helps a user understand

0:19:25.440 --> 0:19:29.040
<v Speaker 4>or a fan understand what just happened. Again, some key

0:19:29.080 --> 0:19:32.639
<v Speaker 4>stats what's upcoming really help us with that storytelling. In

0:19:32.680 --> 0:19:35.280
<v Speaker 4>the past when we have twenty two courts happening at

0:19:35.320 --> 0:19:37.520
<v Speaker 4>a certain time, we would have to pick and choose

0:19:37.560 --> 0:19:40.000
<v Speaker 4>which stories we think or which matches we think are

0:19:40.040 --> 0:19:42.000
<v Speaker 4>going to have the best stories, and that's a really

0:19:42.080 --> 0:19:45.040
<v Speaker 4>hard thing to predict from an editorial perspective. With our

0:19:45.040 --> 0:19:47.400
<v Speaker 4>match reports, now we'll be able to have full coverage

0:19:47.440 --> 0:19:49.399
<v Speaker 4>of every single match during the main draw.

0:19:50.160 --> 0:19:52.760
<v Speaker 2>So of course I want to talk about jeneritive AI.

0:19:53.000 --> 0:19:55.879
<v Speaker 2>How could we not talk about genitive Of course? What

0:19:55.920 --> 0:19:57.240
<v Speaker 2>are you working on with jenitive AI?

0:19:57.640 --> 0:20:00.360
<v Speaker 4>So match reports is the prime example of it. Match

0:20:00.400 --> 0:20:04.159
<v Speaker 4>reports will be completely using Watson next genera of AI technology,

0:20:04.160 --> 0:20:08.200
<v Speaker 4>and really again to us, it's how can we do

0:20:08.240 --> 0:20:12.320
<v Speaker 4>that storytelling at scale? Tennis is such a data rich sport.

0:20:12.600 --> 0:20:14.879
<v Speaker 4>All sports have data, but tennis has a lot of

0:20:14.920 --> 0:20:18.360
<v Speaker 4>shots and different shot types and ball trajectory and live

0:20:18.400 --> 0:20:21.880
<v Speaker 4>scoring data and umpire chair data and crowd and all that.

0:20:22.280 --> 0:20:25.080
<v Speaker 4>Factoring in jenera of AI really helps us take some

0:20:25.119 --> 0:20:29.080
<v Speaker 4>of that structured and unstructured data really one organize it

0:20:29.119 --> 0:20:32.600
<v Speaker 4>in a way, but then help us quickly tell that

0:20:32.720 --> 0:20:35.240
<v Speaker 4>story at scale to all of our fans. And I

0:20:35.280 --> 0:20:37.960
<v Speaker 4>think we're really just starting to scratch at some of

0:20:37.960 --> 0:20:41.280
<v Speaker 4>the capabilities, and we're really excited about where we're being,

0:20:41.320 --> 0:20:43.560
<v Speaker 4>but we also see the opportunity of even how we

0:20:43.600 --> 0:20:45.920
<v Speaker 4>can grow to new fans and new fans around the

0:20:45.960 --> 0:20:48.000
<v Speaker 4>world using jeneral of AI in the future.

0:20:49.320 --> 0:20:53.080
<v Speaker 2>So I'm curious, and you alluded to this a moment ago,

0:20:53.119 --> 0:20:54.720
<v Speaker 2>but I'd like to talk a little bit more about

0:20:54.760 --> 0:20:58.480
<v Speaker 2>it because it seems interesting as a technical problem. Right,

0:20:58.640 --> 0:21:03.920
<v Speaker 2>is the nature of turning tennis matches into stories, which

0:21:04.000 --> 0:21:06.399
<v Speaker 2>is fundamentally what we're talking about here in different ways

0:21:06.400 --> 0:21:11.640
<v Speaker 2>in different media, is about taking both structured data, right

0:21:11.760 --> 0:21:15.679
<v Speaker 2>like the stats who you know, points stats matches, and

0:21:15.840 --> 0:21:19.560
<v Speaker 2>also unstructured data, right like commentary and articles and the

0:21:19.640 --> 0:21:24.159
<v Speaker 2>kind of fuzzier parts of storytelling. And so I'm curious

0:21:24.280 --> 0:21:27.520
<v Speaker 2>how AI kind of helps you manage both the structured

0:21:27.560 --> 0:21:28.680
<v Speaker 2>and the unstructured data.

0:21:29.160 --> 0:21:32.960
<v Speaker 4>Yeah, so I think the structured data is pretty self experimentatory,

0:21:33.280 --> 0:21:35.160
<v Speaker 4>but when you get into the unstructured data and from

0:21:35.200 --> 0:21:37.520
<v Speaker 4>the punditry, that's where you get more of the opinion

0:21:37.600 --> 0:21:40.840
<v Speaker 4>pieces into it. Like a specific player matchup, this player

0:21:40.880 --> 0:21:43.439
<v Speaker 4>always plays well against so and so, or is they

0:21:43.480 --> 0:21:45.440
<v Speaker 4>play always played well at night, or they're a fan

0:21:45.560 --> 0:21:49.320
<v Speaker 4>favorite and the crowd, you know, adrenaline and the crowd

0:21:49.359 --> 0:21:51.439
<v Speaker 4>being behind you can really motivate you to play a

0:21:51.440 --> 0:21:55.080
<v Speaker 4>lot better. So it pulls in all those unstructured pieces

0:21:55.119 --> 0:21:57.640
<v Speaker 4>and helps us really put some more rigor around it

0:21:57.800 --> 0:22:00.160
<v Speaker 4>and help add and enrich our storytelling with it.

0:22:00.560 --> 0:22:04.200
<v Speaker 2>And so I'm curious when you're starting to use generative

0:22:04.200 --> 0:22:07.719
<v Speaker 2>AI over the past few years, like, what were your

0:22:07.760 --> 0:22:08.919
<v Speaker 2>concerns going into that.

0:22:09.320 --> 0:22:13.800
<v Speaker 4>I think our biggest concern is ensuring that one factually

0:22:14.080 --> 0:22:15.680
<v Speaker 4>it is correct, because it's only as good as the

0:22:15.760 --> 0:22:17.800
<v Speaker 4>data you feed in. And how do you really ensure

0:22:17.800 --> 0:22:20.479
<v Speaker 4>that your model's working right and that the output and

0:22:20.520 --> 0:22:23.320
<v Speaker 4>the data you're feeding it matches the output, and how

0:22:23.320 --> 0:22:25.399
<v Speaker 4>do you do that at scale? So we do have

0:22:25.440 --> 0:22:28.640
<v Speaker 4>a lot of human intervention. That's where the IBM consulting team,

0:22:28.720 --> 0:22:31.000
<v Speaker 4>they're on site with us for those full three weeks

0:22:31.040 --> 0:22:34.399
<v Speaker 4>really helping us review everything and we're constantly learning, especially

0:22:34.440 --> 0:22:37.160
<v Speaker 4>early in the tournament. And I would say the other

0:22:37.440 --> 0:22:40.080
<v Speaker 4>big concern, again it goes around to the data, is

0:22:40.359 --> 0:22:43.080
<v Speaker 4>what data do we have available that is trustworthy? So

0:22:43.280 --> 0:22:45.280
<v Speaker 4>you know, we are feel very confident with the data

0:22:45.280 --> 0:22:46.920
<v Speaker 4>that comes off of court, but when we get into

0:22:46.920 --> 0:22:50.640
<v Speaker 4>that unstructured piece. What are the right data sources, how

0:22:50.680 --> 0:22:53.040
<v Speaker 4>do we validate those data sources, and how do we

0:22:53.480 --> 0:22:56.280
<v Speaker 4>ensure that they're accurate? Because if the data that has

0:22:56.320 --> 0:22:58.960
<v Speaker 4>to go in has to be accurate for the output.

0:22:58.920 --> 0:23:01.760
<v Speaker 2>So how do you do that? That's the concern? How

0:23:01.800 --> 0:23:02.439
<v Speaker 2>do you address it?

0:23:02.680 --> 0:23:05.280
<v Speaker 4>Yeah, so I think there's a number of tools that

0:23:05.320 --> 0:23:08.040
<v Speaker 4>we use, all within the Watson X umbrella. We do

0:23:08.119 --> 0:23:10.760
<v Speaker 4>a lot of training with the IBM team, so we

0:23:10.840 --> 0:23:14.520
<v Speaker 4>have to constantly train and retrain that model. I think

0:23:14.560 --> 0:23:17.400
<v Speaker 4>the other piece that we're doing is again as we're

0:23:17.440 --> 0:23:20.200
<v Speaker 4>creating that content, and we have the IBM consulting team

0:23:20.240 --> 0:23:22.760
<v Speaker 4>on site helping us with that, is as we see

0:23:22.760 --> 0:23:25.760
<v Speaker 4>things and we see outputs, it's refeeding that back into

0:23:25.760 --> 0:23:27.560
<v Speaker 4>the model to make it better for the next time.

0:23:27.680 --> 0:23:31.160
<v Speaker 4>So it's a constantly learning process that we're undergoing.

0:23:31.760 --> 0:23:35.520
<v Speaker 2>So I want to talk about scale. Yes, you have

0:23:35.680 --> 0:23:39.440
<v Speaker 2>like what twenty two different courts with matches going all

0:23:39.480 --> 0:23:43.119
<v Speaker 2>at the same time. You're trying to, you know, approximately

0:23:43.480 --> 0:23:46.879
<v Speaker 2>instantly generate summaries of all these matches in something like

0:23:46.960 --> 0:23:51.760
<v Speaker 2>real time. And I'm curious in particular how the IBM

0:23:51.840 --> 0:23:56.040
<v Speaker 2>models you're using the IBM Granite models are helping you scale.

0:23:56.520 --> 0:23:59.280
<v Speaker 4>Yeah. So I think one of the big learnings we

0:23:59.320 --> 0:24:02.720
<v Speaker 4>had with the IBM granted models too is that we're

0:24:02.760 --> 0:24:05.760
<v Speaker 4>able to run it, you know, against last year's tournaments

0:24:05.800 --> 0:24:08.159
<v Speaker 4>and see what the what the expected outputs could be

0:24:08.600 --> 0:24:10.879
<v Speaker 4>and really help train that model heading into the tournament.

0:24:10.880 --> 0:24:13.199
<v Speaker 4>Because as we talked about in the beginning, is we

0:24:13.240 --> 0:24:15.480
<v Speaker 4>can plan, play and plan, but once two players get

0:24:15.520 --> 0:24:18.040
<v Speaker 4>on court, the outcome is unknown. So how do we

0:24:18.119 --> 0:24:20.560
<v Speaker 4>really run it through its paces and really make sure

0:24:20.600 --> 0:24:23.320
<v Speaker 4>that whatever that outcome could be and whatever that scenario is,

0:24:23.320 --> 0:24:26.520
<v Speaker 4>whether it's a fifth set tie break that happens, or

0:24:27.000 --> 0:24:29.400
<v Speaker 4>maybe there's a you know, a fault in the match

0:24:29.640 --> 0:24:32.720
<v Speaker 4>or something that we're not anticipating, that we have that

0:24:32.760 --> 0:24:35.360
<v Speaker 4>accounted for and that the a won't throw off that output.

0:24:35.400 --> 0:24:38.840
<v Speaker 4>So we really try to think through every scenario, which

0:24:39.640 --> 0:24:42.800
<v Speaker 4>is sometimes difficult, right because again live sports is the

0:24:42.880 --> 0:24:44.840
<v Speaker 4>unknown is the unknown, and that's what makes it fun.

0:24:45.160 --> 0:24:47.640
<v Speaker 4>We do spend a lot of time thinking through potential

0:24:47.640 --> 0:24:50.320
<v Speaker 4>scenarios and ensuring that we have the right data sets

0:24:50.359 --> 0:24:52.600
<v Speaker 4>and the model to predict that.

0:24:53.359 --> 0:24:56.680
<v Speaker 2>Tell me about match reports and the generative AI model,

0:24:56.720 --> 0:24:57.399
<v Speaker 2>you're using for that.

0:24:58.080 --> 0:25:00.560
<v Speaker 4>Yeah, so match reports would be new for us this year,

0:25:00.640 --> 0:25:03.160
<v Speaker 4>So we're in testing right now, so we're really excited

0:25:03.160 --> 0:25:05.480
<v Speaker 4>around it. But the model that we'll be able to

0:25:05.560 --> 0:25:08.960
<v Speaker 4>use using Watson X will use a bunch of different

0:25:09.000 --> 0:25:12.000
<v Speaker 4>parts of the suite of tools A meaning that again

0:25:12.040 --> 0:25:14.720
<v Speaker 4>of taking some of that punditry and the unstructured data

0:25:14.720 --> 0:25:18.040
<v Speaker 4>and the editorial spen, take our structured data as well.

0:25:18.359 --> 0:25:21.480
<v Speaker 4>And really what we're working on right now is figuring

0:25:21.520 --> 0:25:24.760
<v Speaker 4>out the right prompts for the AI to really ensure

0:25:25.160 --> 0:25:29.800
<v Speaker 4>that it tells the right structured story, meaning what just happened. Right,

0:25:29.920 --> 0:25:32.840
<v Speaker 4>So our recap is pretty standard. Here's what the data

0:25:32.880 --> 0:25:35.159
<v Speaker 4>is telling us, who won, who lost? How many sets?

0:25:35.160 --> 0:25:35.800
<v Speaker 4>Here's the score?

0:25:35.800 --> 0:25:37.800
<v Speaker 2>The structured data part, that's the easy part.

0:25:38.040 --> 0:25:40.760
<v Speaker 4>Yeah, and then really where it gets exciting is then

0:25:41.040 --> 0:25:44.680
<v Speaker 4>what does this mean? Meaning what's upcoming? So there's all

0:25:44.680 --> 0:25:46.920
<v Speaker 4>these different scenarios when you get into you know, two

0:25:47.000 --> 0:25:49.679
<v Speaker 4>hundred and fifty four players and a large straw. This

0:25:49.760 --> 0:25:52.320
<v Speaker 4>allows us to distill that down and really tell kind

0:25:52.359 --> 0:25:55.280
<v Speaker 4>of what could happen upcoming. The AI helps us do

0:25:55.400 --> 0:25:56.280
<v Speaker 4>that at scale.

0:25:56.520 --> 0:25:58.840
<v Speaker 2>So I want to sort of generalize for a moment

0:25:58.880 --> 0:26:02.399
<v Speaker 2>to talk about kind of you know, broader challenges with

0:26:02.480 --> 0:26:05.400
<v Speaker 2>AI and how you've solved them. You know a lot

0:26:05.440 --> 0:26:10.360
<v Speaker 2>of generative AI pilots fail because the data quality isn't

0:26:10.400 --> 0:26:13.679
<v Speaker 2>high enough, because the risk controls aren't there, and so

0:26:13.760 --> 0:26:17.040
<v Speaker 2>I'm curious how you dealt with those problems and are

0:26:17.080 --> 0:26:17.960
<v Speaker 2>dealing with them.

0:26:18.000 --> 0:26:20.959
<v Speaker 4>Data quality, Again, we feel common with the data that

0:26:21.119 --> 0:26:24.800
<v Speaker 4>is supplied from the US open and from the USTA, right,

0:26:24.920 --> 0:26:27.679
<v Speaker 4>So we have again that's our structure, scoring data and

0:26:27.720 --> 0:26:30.639
<v Speaker 4>all that. I think what we're constantly looking at is

0:26:30.640 --> 0:26:33.120
<v Speaker 4>when we get outside of our known sources and out

0:26:33.119 --> 0:26:35.280
<v Speaker 4>to third parties, is that's where a lot of the

0:26:35.359 --> 0:26:38.760
<v Speaker 4>testing and model work happens. So we pull in different

0:26:38.840 --> 0:26:42.320
<v Speaker 4>data sources and really tried to work through how it

0:26:42.440 --> 0:26:44.600
<v Speaker 4>changes that output. Again, some of that comes down to

0:26:44.720 --> 0:26:47.000
<v Speaker 4>where it's an open model and the transparency that we

0:26:47.160 --> 0:26:49.960
<v Speaker 4>have and the learning that comes behind it. That's where

0:26:50.000 --> 0:26:52.199
<v Speaker 4>a lot of that confidence can come from, and it

0:26:52.200 --> 0:26:55.240
<v Speaker 4>comes from a lot of testing and feeding it more data.

0:26:56.080 --> 0:26:58.480
<v Speaker 4>Your second question was a little bit more around the

0:26:58.520 --> 0:26:59.679
<v Speaker 4>output I believe.

0:26:59.480 --> 0:27:02.679
<v Speaker 2>Right, yeah, and risks, right, So risk I think of

0:27:02.800 --> 0:27:05.560
<v Speaker 2>risk more in terms of output, right, But the obvious

0:27:05.560 --> 0:27:08.640
<v Speaker 2>sphere is like, what if it says something wrong? Yeah,

0:27:08.680 --> 0:27:12.320
<v Speaker 2>inflammatory or whatever like that seems scary.

0:27:12.600 --> 0:27:15.000
<v Speaker 4>Yeah, it definitely is, and it's definitely one of our

0:27:15.040 --> 0:27:17.760
<v Speaker 4>largest concerns when we first took this foray. I would

0:27:17.760 --> 0:27:19.840
<v Speaker 4>say a lot of that comes through our work with

0:27:19.960 --> 0:27:23.560
<v Speaker 4>IBM and IBM consulting team and really ensuring that again

0:27:23.600 --> 0:27:26.359
<v Speaker 4>they're an extension and the partnership there of our team.

0:27:26.800 --> 0:27:29.800
<v Speaker 4>So whenever we are creating let's say it's the Match Report,

0:27:29.880 --> 0:27:32.199
<v Speaker 4>and we're going to be creating these extinct articles for

0:27:32.320 --> 0:27:35.719
<v Speaker 4>every single men's and women's single match that happens, is

0:27:35.760 --> 0:27:38.840
<v Speaker 4>all of those will have manual review and people looking

0:27:38.840 --> 0:27:41.520
<v Speaker 4>through them for accuracy to ensure that the model then

0:27:41.560 --> 0:27:44.040
<v Speaker 4>hallucinat or make up a fact or fill in the

0:27:44.080 --> 0:27:46.600
<v Speaker 4>gaps and things like that. That's the first step. And

0:27:46.640 --> 0:27:49.240
<v Speaker 4>then also when our editorial team goes to publish those

0:27:49.280 --> 0:27:51.680
<v Speaker 4>of the website, they're going to be checking it as well.

0:27:51.720 --> 0:27:54.760
<v Speaker 4>So there are manual interventions throughout that to really check

0:27:54.800 --> 0:27:58.080
<v Speaker 4>that model. But we feel that the ability to do

0:27:58.119 --> 0:28:00.879
<v Speaker 4>it at scale and with us more to that is

0:28:00.920 --> 0:28:03.159
<v Speaker 4>the efficiency problem that we've been looking to solve.

0:28:03.960 --> 0:28:07.159
<v Speaker 2>So the USTA and IBM have been working together on

0:28:07.400 --> 0:28:10.119
<v Speaker 2>digital innovation for like thirty years from you know, the

0:28:10.160 --> 0:28:14.880
<v Speaker 2>first website, yes for the USTA until now. So that's

0:28:14.920 --> 0:28:18.600
<v Speaker 2>the past thirty years. If you look ahead, what's the next.

0:28:18.400 --> 0:28:21.680
<v Speaker 4>Thirty thirty years is a really long time? About A

0:28:21.680 --> 0:28:25.760
<v Speaker 4>three Yeah, I think you know where I get excited,

0:28:25.800 --> 0:28:27.840
<v Speaker 4>and I think I alluded to it in the beginning

0:28:27.880 --> 0:28:30.439
<v Speaker 4>about how I feel like we're just scratching at the surface,

0:28:30.560 --> 0:28:32.720
<v Speaker 4>especially with journat of Ai, and where I see it

0:28:32.760 --> 0:28:35.920
<v Speaker 4>going is there's a lot of different fans out there,

0:28:36.200 --> 0:28:38.080
<v Speaker 4>and we're also very kind in the US open that

0:28:38.080 --> 0:28:40.480
<v Speaker 4>we're a worldwide event and that there's a lot of

0:28:40.480 --> 0:28:44.720
<v Speaker 4>different fans that we're not necessaryly creating content for bespoke,

0:28:44.840 --> 0:28:48.120
<v Speaker 4>meaning in their native language or maybe it's in that

0:28:48.200 --> 0:28:51.160
<v Speaker 4>native players language and things like that. Is where I

0:28:51.200 --> 0:28:54.360
<v Speaker 4>get excited is we've seen immense growth with a Highlights

0:28:54.360 --> 0:28:56.400
<v Speaker 4>and the ability to now do highlights at scale. Is

0:28:56.480 --> 0:29:00.880
<v Speaker 4>the ability for us to start creating content diferent languages,

0:29:01.240 --> 0:29:03.560
<v Speaker 4>maybe covering different parts of the match. So maybe you

0:29:03.600 --> 0:29:06.280
<v Speaker 4>do have that stats junkie who really wants just it's

0:29:06.360 --> 0:29:09.360
<v Speaker 4>the fastest serve and here's the deep insights versus the

0:29:09.480 --> 0:29:12.400
<v Speaker 4>casual fan who's looking for more of the storytelling around

0:29:12.840 --> 0:29:15.120
<v Speaker 4>how a player trains and what leading up to it

0:29:15.280 --> 0:29:17.960
<v Speaker 4>was like and what it means for them afterwards and

0:29:18.040 --> 0:29:20.280
<v Speaker 4>things like that. A lot of that takes a lot

0:29:20.280 --> 0:29:23.280
<v Speaker 4>of time. Now we're able to solve that efficiency problem

0:29:23.360 --> 0:29:25.920
<v Speaker 4>and do it in multiple languages, we can really create

0:29:26.160 --> 0:29:29.440
<v Speaker 4>I want to say, personalized content to a lot more

0:29:29.480 --> 0:29:32.720
<v Speaker 4>fans all around the world, which again helps us grow

0:29:32.720 --> 0:29:34.760
<v Speaker 4>the sport of tennis great.

0:29:35.560 --> 0:29:38.720
<v Speaker 2>So I want to finish with a speed round. Okay,

0:29:38.880 --> 0:29:39.560
<v Speaker 2>are you ready?

0:29:39.680 --> 0:29:40.520
<v Speaker 4>I am ready.

0:29:40.600 --> 0:29:44.080
<v Speaker 2>Okay, first thing that comes to mind, complete this sentence.

0:29:44.720 --> 0:29:46.480
<v Speaker 2>In five years, AI.

0:29:46.280 --> 0:29:49.440
<v Speaker 4>Will transform many parts of the business.

0:29:50.040 --> 0:29:54.320
<v Speaker 2>What is the number one thing that people misunderstand about AI?

0:29:54.840 --> 0:29:59.000
<v Speaker 4>That it's supplemental, not replacing, meaning that it helps it

0:29:59.040 --> 0:30:03.080
<v Speaker 4>with efficiencies, but it doesn't necessarily replace the creativity.

0:30:03.800 --> 0:30:07.520
<v Speaker 2>Right now, what advice would you give yourself ten years

0:30:07.520 --> 0:30:10.400
<v Speaker 2>ago to better prepare you for today?

0:30:11.480 --> 0:30:15.000
<v Speaker 4>I think it would have been, especially now that we're

0:30:15.000 --> 0:30:17.480
<v Speaker 4>able to take so much of that unstructured data and

0:30:18.000 --> 0:30:21.120
<v Speaker 4>pass content that we were created to help tell stories.

0:30:21.720 --> 0:30:24.560
<v Speaker 4>Was to I want to say, archive more of that

0:30:24.640 --> 0:30:26.480
<v Speaker 4>in a way that we could be using that to

0:30:26.560 --> 0:30:30.040
<v Speaker 4>help pull from that now. So you know, we've seen

0:30:30.160 --> 0:30:32.200
<v Speaker 4>kind of a change in the guard from some of

0:30:32.200 --> 0:30:35.080
<v Speaker 4>our start players to now new and up and comers,

0:30:35.160 --> 0:30:37.160
<v Speaker 4>and it would be really fascinating to me if there

0:30:37.200 --> 0:30:39.960
<v Speaker 4>was a way to to cross sections some of that

0:30:40.040 --> 0:30:42.800
<v Speaker 4>and saying like what tra directories are certain up and

0:30:42.840 --> 0:30:46.560
<v Speaker 4>coming players maybe filing from others. So it's more I

0:30:46.600 --> 0:30:48.960
<v Speaker 4>wish we kept more of the content we created.

0:30:48.640 --> 0:30:53.560
<v Speaker 2>Back fave the data exactly. Well are you saving it

0:30:53.640 --> 0:30:54.080
<v Speaker 2>all now?

0:30:54.320 --> 0:30:57.360
<v Speaker 4>Oh? Yeah, one hundred percent learned our lesson? Yes, yes.

0:30:57.960 --> 0:31:00.360
<v Speaker 2>So on the business side of AI, what do you

0:31:00.360 --> 0:31:01.600
<v Speaker 2>think is the next big thing?

0:31:02.240 --> 0:31:05.520
<v Speaker 4>I alluded to it earlier. I think it's personalization and

0:31:05.520 --> 0:31:09.160
<v Speaker 4>getting content that's catered to you at scale, whether you

0:31:09.160 --> 0:31:12.320
<v Speaker 4>know that's across the sports sphere or any type of

0:31:12.360 --> 0:31:16.920
<v Speaker 4>written content or news content. I feel like the ability

0:31:16.960 --> 0:31:19.920
<v Speaker 4>to really get contentated to the type of fan you

0:31:19.960 --> 0:31:22.920
<v Speaker 4>are and the insights you have is where we're all headed.

0:31:23.960 --> 0:31:27.800
<v Speaker 2>And in terms of your non work life, how do

0:31:27.840 --> 0:31:29.480
<v Speaker 2>you use AI day to day?

0:31:29.680 --> 0:31:32.040
<v Speaker 4>It's funny, I was just having this conversation with a

0:31:32.040 --> 0:31:34.280
<v Speaker 4>friend the other day and we were talking about that

0:31:34.960 --> 0:31:38.160
<v Speaker 4>sometimes when you're starting something new, the hardest thing to

0:31:38.240 --> 0:31:40.400
<v Speaker 4>do is you have a blank piece of paper or

0:31:40.400 --> 0:31:43.200
<v Speaker 4>a thought, and how do you get started. Sometimes with

0:31:43.320 --> 0:31:46.600
<v Speaker 4>these generative models, the easiest thing and the best thing

0:31:46.600 --> 0:31:49.280
<v Speaker 4>you can do is it helps you get started. Meaning

0:31:49.320 --> 0:31:51.360
<v Speaker 4>it may not be one hundred percent with that first prompt,

0:31:51.360 --> 0:31:54.240
<v Speaker 4>but it's that efficiency of whether it's an outline for

0:31:54.280 --> 0:31:56.640
<v Speaker 4>a new idea, or it's a marketing brief you have

0:31:56.720 --> 0:31:59.240
<v Speaker 4>to write, or sometimes even if it's an email you

0:31:59.280 --> 0:32:01.720
<v Speaker 4>have to write for personal something and you're not sure

0:32:01.800 --> 0:32:03.840
<v Speaker 4>how to word it the right way. It allows you

0:32:03.920 --> 0:32:06.200
<v Speaker 4>to have a start and then you can edit from there.

0:32:06.240 --> 0:32:09.360
<v Speaker 4>So again going back to my efficiency point, it helps

0:32:09.360 --> 0:32:10.080
<v Speaker 4>you become more.

0:32:09.920 --> 0:32:12.320
<v Speaker 2>Efficient, solves the blank page problem.

0:32:12.480 --> 0:32:13.000
<v Speaker 4>It does.

0:32:14.240 --> 0:32:15.960
<v Speaker 2>Brian, it was great to talk with you. Thank you

0:32:16.000 --> 0:32:16.800
<v Speaker 2>so much for your time.

0:32:16.920 --> 0:32:18.440
<v Speaker 4>Yeah, this was fun. Thanks for having me.

0:32:20.560 --> 0:32:22.800
<v Speaker 3>A huge thanks to Jacob and Brian for the deep

0:32:22.880 --> 0:32:26.959
<v Speaker 3>dive into the cutting edge innovations transforming the game of tennis.

0:32:27.560 --> 0:32:30.480
<v Speaker 3>Brian shed light on how the US opens partnership with

0:32:30.560 --> 0:32:35.800
<v Speaker 3>IBM is harnessing data driven insights to reshape storytelling in sports,

0:32:36.240 --> 0:32:41.760
<v Speaker 3>from AI generated commentary to match reports. As we look ahead,

0:32:41.840 --> 0:32:46.600
<v Speaker 3>I'm excited about the possibilities for personalizing content and reaching

0:32:46.640 --> 0:32:50.680
<v Speaker 3>fans in new ways. The future of AI promises more

0:32:50.760 --> 0:32:59.080
<v Speaker 3>than just efficiency. It's about enhancing fan experiences worldwide. Smart

0:32:59.080 --> 0:33:03.040
<v Speaker 3>Talks with ib produced by Matt Romano, Joey Fishground and

0:33:03.160 --> 0:33:07.640
<v Speaker 3>Jacob Goldstein. We're edited by Lydia gene Kott. Our engineers

0:33:07.680 --> 0:33:11.720
<v Speaker 3>are Sarah Bruguer and Ben Tolliday. Theme song by Gramscow.

0:33:12.680 --> 0:33:15.560
<v Speaker 3>Special thanks to the eight Bar and IBM teams, as

0:33:15.600 --> 0:33:19.080
<v Speaker 3>well as the Pushkin marketing team. Smart Talks with IBM

0:33:19.320 --> 0:33:23.680
<v Speaker 3>is a production of Pushkin Industries and Ruby Studio at iHeartMedia.

0:33:24.200 --> 0:33:28.360
<v Speaker 3>To find more Pushkin podcasts, listen on the iHeartRadio app,

0:33:28.680 --> 0:33:34.400
<v Speaker 3>Apple Podcasts, or wherever you listen to podcasts. I'm Malcolm Gladwell.

0:33:34.720 --> 0:33:38.440
<v Speaker 3>This is a paid advertisement from IBM. The conversations on

0:33:38.480 --> 0:33:54.960
<v Speaker 3>this podcast don't necessarily represent IBM's positions, strategies or opinions.