Daniel Shichman: Mixing Artificial Intelligence (A.I.) with Sports Content (Full Transcript)

Listen on Apple | Spotify | Google

Full Transcript:

Ronen Ainbinder 0:31

Today's snack features the CEO & founder of one of the most exciting technologies in the world of sports.

We're talking about WSC Sports – which is the most prominent Israeli sports technology company.

WSC sports was founded in 2012 in Tel Aviv by a group of friends (one of which is here with us today).

The initial idea was to provide short-form videos for coaches, scouts, and training staff.

However, once they realized the potential, they pivoted to AI-generated highlights for sports broadcasts.

The platform identifies and analyzes every event that occurs in live sports broadcasts. And in real-time, creates and publishes customized short-form videos to a broad set of different networks.

WSC has raised +39 million USD to date, and they work with over 200 leagues and organizations worldwide, such as the NBA, NHL, ESPN, FanDuel, Bleacher Report, and many more.

It's a complete honor to host the CEO and co-founder of the global leader in artificial intelligence-driven sports video technology.

Ladies and gents,

Daniel Shichman!

Daniel Shichman 1:41

Great interesting Great to be here. It's gonna be fun.

Ronen Ainbinder 1:44

It's an honor for me to have you on my show. I've been following WSC for a while now. And man, let's kick it off with an icebreaker straight away, and they want to know what your most-watched movie?

Daniel Shichman 1:56

Let's watch a movie when I was younger. Obviously, it was Ace Ventura pet detective like the first one I watched, I think eight to 10 times or something like that.

Ronen Ainbinder 2:06

And is it like the same one today or you've you have a new one preferred one?

Daniel Shichman 2:11

Unfortunately, I haven't had time to see movies in the last few years. With a company, family, kids, and other things going, I hope to get back to it in a few years. What's yours running?

Ronen Ainbinder 2:21

Mine, I think it's Toy Story. I just watched it repeat as a kid as you finish it, and then you start it again. Yeah, so that was mine. Man. I'm so excited to have you on the show. Plan to ask us about so many different things related to sports and technology to WMC about AI and technology powering the platform. Also, maybe we'll talk about leverage and content and the future of Ott. But I want to kick it off, maybe just by starting to understand who you are in terms of building your curriculum in sports. From listening to one of the other podcasts that you appeared in, I remember that you played basketball; you were part of a league. So maybe tell us a little bit about your curriculum, specifically in sports, and your technology curriculum, like how those two came together to build the dental shipment that we know today.

Daniel Shichman 3:26

So yes, once has been a big part of my life, ever since I was a child, I played basketball in the team. Until I was 18, I was even quite good. And then obviously a huge part of my life was basketball and training and everything. And obviously, as part of it, a huge sports fan that matches almost everything on TV, whether it's volleyball, car racing, golf, like everything out there. And also had another job as a manager in a football manager game on the PC, if you know that. So sports was a very big thing for me. In addition to that, I was a good student. But I was never like one of those wonder kids that built computers or learned how to code when they were nine. But after I finished high school, I went to study electrical engineering, then served in the Army as an engineer and got more experience and exposure on how to build systems and what can be done. And I liked the technology and the engineering side of it. And then I just tried to find a way to combine sports with technology and work with friends. And then, as we mentioned, we try to do something as a hobby. And one of the founders was a coach. I played basketball, so we said, okay, let's do something fun. It was just as a hobby at night to build video scouting software to help coaches learn their opponents' strengths and weaknesses. And we don't know how but it started to be a good tool, which sold to a few premier teams in Europe. And then we got exposure and quality means not only to build a product and sell it but also how to support it and how to get feedback in the industry and how to fine-tune it, and then just think on going to market and, and other things that you just learn by getting into the mud and doing it. And yeah, and after we finished army service, we wanted to stay in sports and not go to the old engineering places that we could have. We started researching the sports industry, and we saw that meteorites are a big thing.

And along with trends that we saw in the industry, with consumer behavior in piracy. All these things that we thought that's a space that was getting into, we got into it, this is our first real job ever since. And we're, and we're learning by making a lot of mistakes. So I think that is how we got shaped by a lot of trial and error in our own life experiences, for good or the bad.

Ronen Ainbinder 5:47

How quickly did the project go from idea to execution when you and hi and your friends agreed on building WSC?

Daniel Shichman 5:58

So if we're talking about the media pieces we do today, I think once we had a general direction about the fact that we want to start creating a platform that knows how to index sports and create content. We had a lot of options, then we started going to the market and ask questions; that was the first thing that we did, we just went and spoke to a lot of sports organizations in Switzerland, in the US in different places to try to understand where they where's the real pain, and what can we solve and what is real, additive and impactful. And then, some of them started to tell us that content creation is important. The personalization is getting there. And they told us that the task was hard, and they didn't think we could make it. And that even made it even more compelling to get into it. And then the product itself, the first version, which is always not good enough, I was ready in a few months already. And then you give it to the first client, which luckily for us was the NBA. And then and then there was like they were an amazing design partner all credits to them because we got real fast feedback and how we need to improve stuff. And we didn't sleep at all for a full year. We're awake at night at Israel time to see how things are working and then waking and then staying up and then in the morning building it and improving it all the time. Almost like that portfolio. And then from that point on, things are just, this is still the mentality, but we keep growing.

Ronen Ainbinder 7:21

Tons of respect to you and to four of your friends, for friends that agreed on building WRC. It's hard to align everyone on the same page and have everyone work in the same direction. Sometimes, founders prefer to either go on their own or just find one co-founder. So I wonder if there was any specific agreement or deal that you guys made, like, I'm going to be in charge of this, you're going to be in charge of that. Like how did you guys do it, too, like agree on working on this and still be very efficient towards it?

Daniel Shichman 8:01

So I think that the first most important thing is that we're very good friends. That's even before everything we're good friends, we get along together, we have fun together, and we do it well. But in addition to being friends with very different people, like if you look at us and talk to us, we say how they are not even close to each other. And everyone is like different and then everyone got greeted naturally to their to different places and want to the technology side want to the operation side ones the business side, and I'm like the kindergarten side. So everyone that everyone is, so that's the easy part. And then from that point on, as everyone has, has their own domain, they're in charge of it, and they know how to run it. And we all trust each other. Like that, we'll do the best we can even if we make a lot of mistakes. And so it sounds too ideal, but it is, at least for the last 10 years or so; yeah, it's been fun. And I'm blessed to have these guys in the race.

Ronen Ainbinder 9:00

That's amazing. Danny, let's talk about the technology of WSC. Specifically, let's talk about the elements that make up WSC. I remember reading there was audio recognition. There are also video real-time video analyses and specific algorithms. But can you mention just over the top, what are the main components of the technology that make up AI, but AI is also built off of some components? So maybe just tell us real quick what are like the components that make up WSC?

Daniel Shichman 9:36

Yeah, so the way before that, the way the platform works is that we ingest a live stream. So the same one that you watch as a fan and watch the game. Our platform gets it and tries to analyze it and then try to understand everything that happens within the game to identify every action, whether it's a dunk, a goal, or a touchdown, the exact time it happened, and the perfect start and end time are. So if there were three passes in to make sure that's included in the clip, or if there's a close-up of the player that just scored the basket to make sure that's included. We don't cut the commentator, and we also add ratings on how interesting or exciting The play is the way it works. And that your question is, as we get the stream, we try to analyze it in the game in three different layers. The first one is real video analysis. And we have per sport different things and mechanisms that we're looking at. Whether it's identifying specific locations of players or identifying the specific action like a golf swing, we have a neural network that knows how to identify that a shot in basketball, a pass in football, and things like that. We also identify the different scenes when and many things, so everything we can extract from the video footage itself. The second layer is the audio layer analysis. And then it's what is being said, but how the audio behaves. So if the commentator speaks fast, or starts to shout, or the crowd goes crazy, or your specific sounds per sport, that is helping us to also get indications and identify the things that are happening. And the third layer is data that we're just ingesting play-by-play data that is out there; many companies are providing it. And so once we do video analysis, audio analysis and data analysis, combining everything into like, the black box and from that point on try to understand exactly as I've mentioned, what happened, when who did it, what is the start of it? What is the end? How interesting it is. And once we have everything broken down, we analyze more than 100 streams per year. And we have all of the little segments, then comes the second part of that technology where users, the leagues, the broadcasters, can go and set rules and create triggers to what content will be created and where it's going be sent to. So they can say okay, they have come over the game, creating a game highlight. So far, in three different versions, the first one will be one minute vertically to the app, the second will be three minutes to the website, the third will be 10 minutes with a sponsor to YouTube. And the system automatically starts to generate that content and customize it for every platform, every use case. And basically, that's it.

Ronen Ainbinder 12:03

I wanna talk about the vulnerability of the technology or probably the weakness. And here's, here's what I'm thinking. And you tell me, what's your experience with this word. For instance, if you're analyzing audio and the crowd noise, what happens when you don't have a crowd noise? And like, for instance, during COVID, during the bubble, in the NBA, there was no crowd. So how exactly? So let's talk first about how vulnerable is the code? Or how, how precise, how accurate Can you predict things happening in the field with small changes in the inputs. And let's also talk about how the pandemic shifted your code on a fundamental level. So first, talk about vulnerability; how vulnerable is the code?

Daniel Shichman 12:53

So basically, when you go and train a neural network, the, what's called, the traditional way to do it is you collect a lot of data, a lot of samples of what you want the system to learn, it's like a four-year-old kid, you show them something, a lot of times, he will learn that this is a horse, or this is a dog or whatever. So it's the same way you show a lot of samples. And as you have more variety, and more things that are different angles, different lenses, different whatever the system learns better. And obviously, if you come and show it something it has never seen before, the system will not identify it if you show a new angle. So that's, and that's a given thing the same way a person will say something totally new, they will not identify the kid, or grow up. But as the system now analyzes hundreds and hundreds of 1000s of events of every type in every sport, the system is now very mature and identifies a lot of things. In addition to that, there's no AI system. We have an internal confidence system. Whenever the system is not certain of something, I think it is in a basket. Still, it is not certain it provides a flag, and then we know that we need to look at it when we play, identify when to take a second look, and things like that. So that in terms of the vulnerability, and then in terms of the pandemic, what you've asked. So I think the pandemic, like the audio, is obviously one of the indicators. And it helps us to identify what things have happened and how interesting now how exciting they are. But as I've said before, there are other indicators. So the lack of a crying nose might hurt the confidence, as I've mentioned earlier, because it's an indicator that might not exist but also might hurt how exciting the place is. But then on the other end, if you are a fan, if you watch something and if I let you watch a game with fans, and then without fans, you'll see the exact same moment. And if I'll ask you which goal is better, you'll say that the one with the other track is more exciting as a fan. So this is even still the case. So it also matches his reality, the reality, and how much we're excited about sports. And only during the pandemic did we understand how important the crowd noise is to the experience of watching a game where you're sick of sitting at home because you create the highlights. There's no crowd, and then it's noticeable. But if we watch it with crowds and other games, it's a totally different ballgame, so it's worked well on that.

Ronen Ainbinder 15:07

And there was anything that you guys had to add to the code or how to change in the technology in the fundamental level because of COVID because of these new learnings that this new technology this technology didn't experience before was there something that you have to tell it okay so learn that if there's no crowd noise maybe look more or put more attention on the commentator or on this specific hype around the play during like for the players or stuff was there anything specific or not in a very fundamental level?

Daniel Shichman 15:39

No, not at a very fundamental level like the system kept running. It is nothing, luckily for us, nothing special on our end.

Ronen Ainbinder 15:47

Amazing, man. So, this technology blows me away because of all the things it can do on its own. Okay, it's amazing to see how you created a platform that can work for professional teams, leagues, and organizations worldwide. And it can do so many things for them that it can save so much time at the same time for them, and you don't have to, like, be there for it. So I call that leverage and the power of technology and code. But at what moment, did you realize like, or you understood this concept of leverage and the power of like, what you were doing and how much time you were saving organizations by, the technology and the content by particularly in sports, like when what was the moment when you were you realize this was like 1000 times more or better than you ever thought it would be?

Daniel Shichman 16:47

Yeah, it's a good point; I'm not sure there's a specific point in time when we started to pitch the idea that it was always about how you can do more with your content. Until today, you had an editor creating one version. And that's it. But in today's world, or Tomorrow's World, back then in 2013, it won't be enough because people on Instagram because Instagram people on social media are expecting different experiences on the website from different app variations, a different lens, I'm, I'm a fan of a specific theme, I want one on the watch for the content of that theme. And my friend is something else. So we always understood the importance of scale; we mentioned in this scale, to be able to scale and create any type of content for any type of experience, that's something that we put a lot of emphasis on. Obviously, in the last few years, we just see an explosion in the number of platforms out there, and it keeps growing. And I think that's the thing that we're amazed by just to say that if back then was okay, social media would be short content, YouTube a bit longer, and an app would be vertical. And that's it. Now there are TikTok, Instagram, and Snapchat, and all the sponsors are asking for activations and betting and fantasy. All of a sudden, you have so many experiences and so many touchpoints. And every touchpoint requires a different experience to be relevant to the specific fan. And I think this is the fun thing because of creating the relevant content for every platform I've experienced. That's amazing. And we keep being amazed by the league's creativity and what they're doing, how they activate the content and the importance they see with it. So I think it's something that is evolving all the time. And we were amazed, surprised, and happy that it keeps growing all the time.

Ronen Ainbinder 18:27

So how exactly do you grow from here? I understand that you have already worked with a lot of leagues. So probably adding a new one has no marginal cost of replication. One of the most fascinating things about your platform is that you can just adapt it to their needs. But I wonder if, for instance, there's growth that it's harder to tap, meaning if there's probably sports that are harder to analyze, through your technology, I'm just gonna throw one example out there. Still, I don't know if it's like the same thing in snowboarding, or for a specific Olympic sport, like table tennis, where it's like different dynamics and things; things are happening differently. I wonder if it's that type of growth, meaning it takes a sport that it's harder to capture through technology, it's harder to grow to those sports and then to the leagues and organizations related to them? Or what do you think it's the route or the path that you're trying to take in terms of growth? What is the most challenging side of growth? And what is your plan?

Daniel Shichman 19:37

So it's an amazing question because that's what keeps us busy all the time, what's next for WSC, and how we keep growing. So when we look at growth and the next step, the first one is, as you said, keep doing what we're doing, working with more leagues, more sports, more geographies, improving the product, offering more layers and packages, and continuing to do that. And specific sports obviously are: Are there some of ours more similar to others, and we're trying to see what we can adapt and extract from each and every sport and continue on that road and just adding more logos and making our clients happy, and with what we're doing. And the second route, which is looking at new verticals in sports, and where else content can be used. And when we think we can add value. And we see a lot of places that now we think that short form and real-time content can be very relevant and might be sports betting, which is a new vertical, which is going and Ott, which is a variable that is growing, and user and vertical archive content is growing in the NFT. So there are many, and again, as I've seen for every touchpoint, every place a fan consumes sports, every fan platform, a fan spends time, and we think that can be relevant. And when it's adapting the technology to the right business case and why there should be content there. So these are the things that we're going to grow into and basically be the engine that creates relevant content everywhere for fans,

Ronen Ainbinder 20:59

it's insane to hear you speak with such passion and such clarity on those things. Daniel, I'm fascinated by everything that you're telling us. You talked a little bit about NFT's and sports betting? Would you mind talking a little bit more about the NBA Top Shot, the technology that creates specific highlights and then sells them as NFT's? And it's exploring this new avenue. Have you seen other use cases related to this new trend? And would you like to say that is the way sports are moving now? Or do you have any other specific idea and in terms of content, of course, of what fans are not teams but what fans are interested in receiving more of?

Daniel Shichman 21:54

I think two of the trends are to mention both betting and NFT's. We can touch both of them shortly. So betting, especially in the states and see, like the betting operators, like the Federals and DraftKings of the world, are becoming media companies in creating podcasts and TV channels and trying to engage fans not just like them place a bet. But be an app or destination, you open a few times a day even to catch up on sports. And the same way applies the other way around; we'll see media companies trying to involve betting activities within the broadcast and everything. And our thesis is saying, okay, we think that short-form content can be an amazing driver, both for user acquisition because you might tell the user, hey, this just happened, it's interesting, here are the new orders, or here's an insight about it or anything now come here to catch up and maybe place a bet, or just provide a better experience that as if you placed a bet on something, you want to be up there and see what's going on even without spending three hours just watching the game. So we think there are a lot of product lines that are now some of them already launched. And some of them are now in the making to make the experience much better. And in terms of NFT, we see it's a trend. And you mentioned the NBA culture, which was about collectibles and archive content. But you'll also see other companies that some of them are using, like creating assets and using in fantasy sports. And just the concept of owning digital assets is just growing around the world, not only sports is just growing. The trend is out there; I'm sure there will be a lot of companies that will be very innovative and create a lot of experience. So I think if the first stage was about just owning something, I think the second part will be how you use it, and how the fact that you own it will benefit you. If it's part of the game, part of a competition, part of something else, that will be even more, and I think that's where everything is going. And again, I think the fact that you can own something that just happened will be impactful. But we'll see. We'll see. So hopefully,

Ronen Ainbinder 23:46

Lastly, I want to talk about how you came up, or what exactly is the business model of WSC? Is it the amount of time that any of the content they own goes through your algorithm and gets spread out and distributed to different platforms? Or are you charging differently for each one of the leagues and teams that work with you? Or how exactly do you go about thinking, offering the service that is software as a service to all of your customers?

Daniel Shichman 24:22

So the company, as you said in the letter, it's a software service business. So basically, it's a yearly or multi-year license fee that the league's broadcasters teams usually pay us because the few variations or parameters that define the price one is the number of strains that we need to analyze, you know the one they can have 200 games. In comparison, the other can have 1500 games, and obviously, we need to do a lot of computing power. The second is that all the use cases they're using the platform can be just to social media and websites. Still, it can be enhanced to Google search and sponsors and app integration, not many other use cases that will keep growing. The third one is just the product packaging. They can just buy specific modules or just add more and more modules as they move on. And then, once we understand the scope, it's your yearly license that usually just grows over time. So that's on the main product line that we discuss. And then additional ones that are more about user acquisition and user activation are usually with the marketing departments. And then these departments are, are used to paying for activation, pay per click per sign up a prescription, whatever. So that's that, but the main core product is about subscription fees.

Ronen Ainbinder 25:34

And you see WSC expanding to different industries not related to sports.

Daniel Shichman 25:41

So at the moment, we're focused on sports; we're sports fans and want to have fun with what we do. B, I think there is much room to grow in the sports world; there are other verticals and other things that we didn't touch. And at the end of the day, if you look at the content worldwide, sport is the most important and expensive piece of content out there. The live aspect is super important. And I think again, there's a lot of room for growth. So at the moment, for the short and midterm, we're still very focused on sports building the best sports platform out there and

Ronen Ainbinder 26:12

I'm looking forward to everything we've been discussing here is super exciting, fascinating, especially hearing it from you firsthand. Daniel, I can't leave without asking you a more personal question. What has been the best day of your life

Daniel Shichman 26:26

the best day of my life? Wow. I have no answer to that and a lot of good days.

Ronen Ainbinder 26:33

The best day, the best day for WSC?

Daniel Shichman 26:35

The best day for WSC, maybe when signing the first fundraising, was a hard process. And maybe that I can. I think that maybe but we had so many apps and so much fun time that every year is so different like we double every year and everyone it's the challenges, and we're having fun. And so yeah, we just keep enjoying.

Ronen Ainbinder 26:57

Well, Daniel, I can't thank you enough for coming to the halftime snacks and snacking with all the snackers and me about everything related to you your stories, sports technology, WSC, the path that you guys intend taking the technology, the elements, the strength, the vulnerabilities, and yeah, everything, everything related to it. It has been so much fun, and I've learned so much, so I can't thank you enough. Hope we can do this again in the future. And you're always welcome to the Halftime Snacks, Daniel!