Matthew Candy, Global Managing Partner, Generative AI, IBM Consulting Chris Clements, Digital Products Lead, Wimbledon Maggie Murphy, Chief Executive Officer, Lewes FC Moderator: Ellie Austin, FORTUNE
Category
🤖
TechTranscript
00:00Matt, Chris and Maggie, thank you so much for being here.
00:02Now, Matt and Chris, I think we should just explain
00:04how you two work together.
00:05So IBM is the digital partner of Wimbledon,
00:08which is why we have the two of you on stage together,
00:10just so there's no surprises.
00:12Chris, I want to start with you first.
00:15So AI is being used in various ways at Wimbledon.
00:19Could you maybe highlight a couple of the cases
00:21that you think have been most transformative to date?
00:24Yeah, absolutely.
00:25So yeah, like you said, there's a number of ways
00:27that we've been utilizing AI.
00:29I think the first one is around distilling data
00:31and being able to deliver engaging fan experiences
00:34at the back of that.
00:35Now, when I talk about that, it all starts with the fan,
00:37understanding their needs, understanding the pain points,
00:40and understanding what we can serve them
00:42to get them more engaged with Wimbledon
00:43and get them more engaged with the tennis.
00:46So what we've done over the years
00:49is we've done a number of different initiatives,
00:50starting with our power index.
00:52So what this was was a dynamic index,
00:55which enabled users to be able to see
00:59how players are progressing in terms of their form
01:02as you moved on through the tournament.
01:04So something a lot more dynamic than existing rankings.
01:06So users could understand exactly
01:09who they should be looking out for
01:10in terms of the big matches that are coming out.
01:13That then fed through to our match insights,
01:16which are a percentage likelihood to win for each player
01:20on a match that's coming up.
01:22So for a fan, they can see where the big matches are,
01:26who's likely to play, especially for those fans
01:28that are perhaps less engaged,
01:30that don't follow the tour year round.
01:32And where are they consuming this?
01:34So that's on our digital platforms,
01:35as well as on social media.
01:37And then this year, or in 2023,
01:42we took that a step further with our draw analysis.
01:45So what that was was an analysis
01:47of how difficult or favorable a player's draw was
01:50from the very start.
01:51So that enabled a user to be able to select a player
01:54and see what their path to the final looked like,
01:57so they could see where the big matchups were coming,
01:59where they were going to have
02:00a slightly more challenging matchup.
02:02And then they could see that
02:03in order to be able to fully understand
02:05how the draw was likely to play out.
02:07And I think we've got some video footage,
02:08which I'd love to run, showing how this interface works.
02:12And you can also see it down here.
02:13Maybe Chris and Matt,
02:14you could talk us through what we're seeing here.
02:19Yeah.
02:20I mean, tennis, we're seeing tennis.
02:22This is a representation of the tennis
02:24and all of the data points which we're utilizing
02:28in order to come up with some of those stats.
02:29So tennis is an extremely data-rich sport,
02:32but it's not always the easiest
02:33for your average fan to understand.
02:36So what we're doing is utilizing AI
02:38to be able to take all of those data points
02:40and distill that into something which is easy
02:43for anyone to understand
02:44so that they can fully engage with the event.
02:46And Matt, who is the demographic that you're targeting here?
02:48Because it doesn't seem like,
02:50if I think about my parents who love Wimbledon,
02:52I can't imagine they're going to be tuning into the app
02:54to engage with all this data.
02:57So I think it's aimed at all demographics.
02:59So only a small number of people
03:00get to experience physically the Wimbledon,
03:02the amazing Wimbledon environment each year.
03:05So this is all about extending the reach,
03:07the richness of the brand,
03:09and just the beauty of the game and sport.
03:11And so that second screen experience
03:13that people can access through the iPad app, iOS, Android,
03:17through the web platforms acts as an amazing companion,
03:19effectively, to the live broadcast.
03:21And so we end up with all types of demographics
03:24accessing that in order to drive a deeper level
03:28of engagement with the game and with the sport.
03:30Maggie, I'm going to come to you in one second.
03:32Bearing all of this in mind,
03:34as we integrate AI further into the viewing experience,
03:37does what it means to be a sports fan change?
03:41So does what it mean?
03:44I don't think it does.
03:45I think it allows us to actually
03:47make the most of people's time.
03:49And I think if we all think about us in the room,
03:50everybody's chosen to spend time here today,
03:53and time is probably the most valuable commodity
03:55that we've got.
03:56And so if we're able to create experiences
03:59that allow people to personalize the use of that time
04:01and be able to access the information that they want,
04:04and certainly being able to create
04:06personalized content, et cetera,
04:08I think the technology provides an amazing way
04:10to make the most of people's time.
04:12Maggie, let's bring you in.
04:13So in addition to being CEO of Lewis Football Club,
04:16last year you were part of a really interesting initiative.
04:19You were part of a team that created an AI campaign
04:23for a hypothetical female FIFA presidential candidate
04:27to illustrate the fact that there are very few women
04:30at the top levels of football.
04:32And before we ask you about it,
04:34I want to show a clip of the campaign,
04:36and our candidate is called Hope Sogne.
04:40My name is Hope Sogne,
04:42and I am the first female candidate
04:44to stand for FIFA president.
04:46It might seem like a thing of fairytale or fantasy,
04:50for football to be governed by a woman,
04:51but I believe it can one day be our reality.
04:54History has shown us how the power of collective voices
04:57can dismantle broken systems built to protect the few.
05:02Okay, so Maggie, first of all,
05:04Hope looks incredibly realistic.
05:06And secondly, why did you decide
05:08to get involved with this project?
05:10Yeah, well, I mean, first of all,
05:11I don't know anything about AI,
05:12so I'm probably in the wrong room.
05:14I do know a fair amount about people,
05:17and unfortunately, I wish that Hope didn't have to exist,
05:21but she exists because of human-made challenges,
05:23which is the fact that it's very, very rare
05:25for female candidates to get elected
05:28in football governance.
05:29You might find there are plenty of very, very good women
05:32that are working in football,
05:34but when it comes to elections,
05:35which is generally how football is governed,
05:39they don't always get elected.
05:40It's almost like their whole woman-ness
05:42is a little bit too radical.
05:44And so it was, we came together
05:47and decided to create a candidate
05:49and fed this person with the views, the insight,
05:54the research, the data from a whole bunch of women
05:56around the world who are working in football
05:58to create this person.
06:00Hopefully, who someone like Gianni Infantino
06:02could actually ask questions to
06:03to maybe get a woman's perception on a particular challenge.
06:06And the challenges are real.
06:08It's a case of the fact
06:09that take the Ballon d'Or ceremony last year,
06:13one of the biggest award ceremonies in the world.
06:14You hardly have any women
06:16that are in the decision-making roles.
06:18And so it was put during a Women's International Calendar.
06:20So all the top stars
06:21were actually playing football that night.
06:23And then you also get really specific issues
06:25around a number of rape allegations
06:27where female players are actually going through a system,
06:31but the people that are perpetuating the rape,
06:33for example, their own presidents
06:34of the football associations,
06:35are the people that they are required
06:36to report the allegations to.
06:38And so Hope was kind of trying to create a system
06:41where we can feed in all of these ideas
06:45and have a candidate that can actually speak to
06:47and represent women that if they spoke up
06:49in some circumstances,
06:51they could actually face retribution.
06:53Did Hope have any impact on changing the game?
06:56Yeah, well, there was certainly a lot of buzz.
06:58And I think that's what we wanted to do.
06:59We wanted to start a conversation.
07:01I don't really want Hope to be elected.
07:03I'd rather there be a woman out there that is elected.
07:05And that is the point.
07:06It was just to show that there are these situations
07:08around the world where women are not getting,
07:09there's that kind of glass ceiling that's still in place.
07:12So we had plenty of reactions,
07:13plenty of discussions, plenty of conversations,
07:15and even some conferences thought about
07:18actually profiling her and platforming her
07:20and giving her the stage.
07:21So that's exciting.
07:23We missed a trick not having Hope today to us,
07:25but we're thrilled to have you.
07:27Matt, so in addition to working with Wimbledon,
07:30IBM works with Seville Football Club,
07:32with the Golf Masters, with the US Open.
07:34I have a question about how you decide
07:37what use cases to pursue.
07:38So I saw a clip on Twitter or X
07:41of what was AI-enabled commentary at the Golf Masters.
07:45And some people were loving it.
07:46And some people were saying,
07:47if this is what it sounds like, I'm pressing mute.
07:50How do you make sure that viewers want the new technologies
07:54that you're introducing
07:55and that it's not just innovation for innovation's sake?
07:58Yeah, so, I mean, we've just,
08:00obviously, Masters concluded yesterday.
08:01It did.
08:02Been partner with them, I think, for over 25 years now.
08:05And so this year, we had spoken voice commentary
08:09for every shot, for every hole, for every player
08:12in both English and Spanish.
08:13And so the key objective there,
08:15and also similarly, if you look at what we're doing
08:17with Wimbledon and the US Open,
08:19is to basically be more inclusive.
08:21And so if we can provide, effectively,
08:23the ability for people in their natural language
08:26to be able to access deep and rich commentary
08:29around the game that they're watching,
08:32we've done similarly with Wimbledon
08:34around the highlight clips and the highlight reels
08:36that we generated last year
08:37and layering a natural language
08:39spoken voice commentary generated.
08:40Would it ever sound full of emotion
08:42and excited about the game?
08:43Yeah, no.
08:44So I think there's a lot of work that we've done
08:46taking open source large language models
08:48and the small model and training it specifically,
08:51and not just on language,
08:52because this isn't a language translation issue.
08:54This is around picking up the nuances of the game
08:56and expressing, effectively, what people are seeing
08:59in a way in which it's going to build emotion.
09:01And certainly, we've started with English and Spanish
09:04as two of the big constituents of viewing groups,
09:06but we absolutely see this as something
09:08that will open up to other languages beyond.
09:11But I think it's important to say that, actually,
09:13what you're not aiming to do here
09:15is replace the human commentators.
09:17Actually, what we're looking to do
09:18is find ways to supplement them.
09:19So for instance, the highlights clip.
09:22The highlights clips of Wimbledon last year,
09:24we had generative AI commentary
09:26that was overlaid onto them.
09:28What that was offering was context
09:29as to the clip that you were watching.
09:31So it's not about trying to bring out
09:33that human emotion.
09:33It's not John McEnroe screaming
09:35at how exciting the point is.
09:37What it is is just providing
09:38that extra little bit of context,
09:40which enables you to understand
09:41what you're seeing and why it's important.
09:44So it's about finding the right use case for it.
09:47It's not necessarily about replacing
09:49or trying to come up and shoehorn these
09:54into places where they're not necessarily needed.
09:56It's about finding where the gaps are
09:58in the fan experience,
09:59and then we can look at the technology
10:01and whether or not the technology can work
10:03to alleviate some of those challenges
10:05that fans are having.
10:07Sorry.
10:08I was going to say,
10:09and this point of augmentation is really important
10:09because, you know, and I think if I look
10:12at the video generated highlight clips
10:14that we produce,
10:15that's not to replace content producers,
10:17but it's effectively to allow the AI
10:19to take a decision looking at crowd noise
10:21and player facial gestures and fist bumps
10:24to auto curate what we believe
10:26is the most exciting points of the game,
10:28and then allowing a content editor
10:29to push that highlight reel
10:31and reduce the cycle time
10:32for getting stuff out there, basically,
10:33and allow more content and a quicker cycle time.
10:36And Chris, Wimbledon's a very unique example
10:39because it's a tournament that's so steeped in heritage,
10:41and so the balance between preserving that culture
10:43and innovation is a particularly unique one.
10:46I know that the ATP has announced
10:48that it's going to be using electronic line judges
10:50across its tour as of next year, I think.
10:53Wimbledon obviously has in-person line judges,
10:56and I wonder, has using AI
10:58to make judgment calls on court
11:00been discussed for the future of Wimbledon?
11:02So for this year,
11:04we are planning on keeping human line judges.
11:08Obviously, with any new and emerging technology,
11:11we'll be looking at how that progresses.
11:13We work very closely with the other tennis bodies,
11:15and we'll be learning from their experience,
11:17but our decision will be down
11:18to what we think provides the best experience
11:21both for players and for fans.
11:22So it could be in the future.
11:24No comment.
11:25Okay.
11:27Maggie, aside from hope,
11:30how else can AI be used
11:34to bring more diversity to the game?
11:36Oh, it's funny.
11:37I was thinking about how AI could be used in general.
11:41I mean, I run a fan-owned club.
11:43You know, we have 2,400, 2,500 owners
11:46in 41 different countries,
11:48and I think, you know, we're all very aware
11:50as non-AI specialists about targeting
11:53and segmenting your audiences,
11:54but it does make me think,
11:55how about if we were able to segment to the point
11:57that rather than getting a monthly or a weekly newsletter,
12:01you're actually getting a video from me?
12:03So I was telling that person exactly what they want to hear
12:06because we're able to understand exactly what they want
12:08so that they're actually getting a video in their inbox
12:11where I'm saying, hey, Dave, how are you doing?
12:13You saw the results last week,
12:14and it actually could personalize the experience.
12:17The real challenge in all of this is
12:19football's about people.
12:20It's about that human connection,
12:23and I would worry that you would lose some of the value
12:27of me actually writing an email to somebody
12:30if they are receiving a monthly video from me
12:34that gives them what they know they want to hear,
12:36but it's not actually me saying it.
12:38So I think that's the thing
12:38that we have to toe the line on a little bit.
12:43Matt, I'm gonna ask you the next question,
12:45then we're gonna open it up to the floor,
12:47so get ready if you have anything to ask.
12:49I'm interested in IBM's partnership with CIVIA FC,
12:53and specifically, I know that,
12:56similarly to what we had Dr. Beale talking about,
12:58the use of AI to enhance scouting.
13:00A challenge there is privacy
13:03and the sharing of personal biometric data
13:05that belongs to players.
13:06What kind of guardrails or regulations do you have in mind
13:09when you're thinking about that particular use?
13:11Yeah, so I guess for all of the work
13:13that we do in this space,
13:14so we have a platform that we've built in IBM
13:17called What's Next, which is our generative AI technology.
13:20As part of that, there's a governance platform and toolkit,
13:22and so where we put these models into use,
13:25that governance toolkit allows you the ability
13:27to monitor, manage for bias, drift, explainability,
13:31observability, and live use.
13:33You think about all of the regulation
13:34that's coming in around the world
13:35and making sure that you're compliant with all of that.
13:38And so certainly, if I look at the work
13:40we're doing with CIVIA Football Club,
13:42so this is a product that we built for them
13:44called Scout Advisor.
13:45It's powered by our AI platform,
13:48and basically taking a lot of the quant and qual data
13:50that they've got around potential players,
13:53about 200,000 player reports that they have,
13:55and basically producing a very simple digital product
13:58to allow their 20-plus scouts that they've got
14:01to be able to access that information,
14:02to get simple summaries of that information
14:04and the data that's there to help them take better decisions.
14:07So again, it's not to replace them,
14:09but it's basically to help them sift through
14:11the huge volumes of data that they've got.
14:13I mean, these models are amazing at summarizing,
14:16at being able to chat with information,
14:18at being able to classify information.
14:19And so when you think about generative AI,
14:22there's many uses of it in a non-generative way.
14:24And so using this type of technology
14:27to basically help get these insights and gems
14:30from the data there to allow them to find talent.
14:33Does anyone have any questions for the panel?
14:35Yes, we've got one right here at the front.
14:37Please tell us your name and your company.
14:39Yes, hello, Jamie T. Van from Microsoft.
14:42Few years ago, I went to the League of Legends
14:45World Championship in the sport
14:47my children watch most is Fortnite.
14:51This seems like there's a lot of opportunity for AI,
14:53particularly in the e-sports domain.
14:54I was wondering if you could say anything about that.
14:59Well, yeah, so actually, funnily enough,
15:02Wimbledon entered Fortnite last year with Race to Wimbledon,
15:06which was a mini game,
15:07which we launched on Fortnite last year.
15:10I think there's definitely a lot of potential
15:14in that e-sports domain.
15:15It's very interesting because what we're trying to match up
15:20is, as Ellie mentioned before, our heritage
15:24and the fact that Wimbledon is a brand that people know
15:26and they associate with certain things,
15:28with the fact that we're trying to move forwards
15:31and we're trying to deliver more engaging experiences
15:33through a variety of different ways and means.
15:35And at the center of all of that is the human element of it,
15:39is the fact that the remarkable feats of sports people
15:42are remarkable because they are human.
15:44And you can't quite believe it at times, but they are.
15:47In the e-sports domain,
15:49potentially there is a little bit more leeway
15:50to be able to be a bit more expansive in some of that.
15:53And I think it's maybe something that sits
15:54a little bit more naturally there.
15:58I just want to finish with a quick-fire question
16:00for each of you.
16:01So we're obviously in a room of people
16:03who largely work in the business world.
16:05What lessons from the application of AI in sport,
16:09in your experience, would you like everyone here today
16:12to take to their respective sectors?
16:13Maggie, let's start with you.
16:15I mean, number one, for us, when we were designing Hope,
16:18you had to have proper human input.
16:20So even though the brilliant creative agency Dark Horse
16:23is with the ones that put it together,
16:24they needed me and they needed my database
16:28of incredible women to be able to feed it in.
16:30And if we are not being diverse
16:32in the way that we create AI,
16:34we're going to lose out a lot of opportunities,
16:36but also sideline a lot of people as well.
16:40Yeah, I think it's about being led by your users
16:44and your customers, not being led by the technology,
16:46and also knowing the limitations of the technology.
16:49So in our instance, for instance,
16:51we worked specifically with Watson X,
16:54which is the IBM platform,
16:56because it can be specifically trained in our domain.
16:58So we can guarantee the quality of the outputs.
17:01It's going to talk like you would expect Wimbledon
17:03to talk, for instance.
17:05So definitely be led by your users
17:07and know the limitations of the technology.
17:10And then I think, look, I guess a lot of the conversation
17:13around AI and generative AI at the moment
17:15is on productivity, right?
17:17And people looking at how this technology
17:19can be used to drive productivity
17:20and processes within the enterprise.
17:22So HR and procurement, supply chain, customer service,
17:26and it absolutely can,
17:26and it's absolutely going to have a profound impact
17:28in those areas of business.
17:30But I think what the sports world demonstrates in spades
17:35is the use of this technology
17:37to drive new types of products and services
17:39and a deeper engagement with your customers and fans.
17:42And so I think there's a lot that business,
17:44every industry can learn
17:45about the way in which the sports industry
17:47drives really deep personalized engagement
17:49through using AI data and analytics.
17:52Matt, Chris, Maggie, thank you so much for joining us today.
17:55Thank you.
17:56Thank you.
17:57Thank you.