Пулс
Спортрадар: нова парадигма у спортској потрошњи
By Вилијам - 4. фебруар 2025
Exhibiting the advancements being made in its Computer Vision and AI technology suite, Sportradar upped the ante at ICE 2025 with 3×3 basketball matches live on the stand.
The level of complexity is on a different scale to the then-impressive table tennis demo at the ExCeL two years ago, demonstrative of how sophisticated the new technology has become in a relatively short span of time.
Crammed cheek by jowl on the booth watching a match take place, one screen shows a real-time technical visualisation of the action, whilst a second screen demonstrates the capabilities of Sportradar’s 4Sight Streaming technology. Integrating AI and machine learning with Computer Vision, 4Sight generates real-time 3D animations and graphics of the players. At the end of each match, 4Sight creates a match narrative using different metrics such as shot positions that are divided into successful and unsuccessful.
As basketball shoes squeak on the court in front of us, G3 speaks to the man responsible for leading the implementation of AI and computer vision technologies within Sportradar’s vast portfolio of products and services. Luka Pataky, Senior Vice President of Automated Content, explains how the technology engages users ultimately driving increased betting activity.
ЗАКУЦАВАЊЕ
Since those live table tennis demonstrations, Sportradar’s Computer Vision has been rolled out across various sports, including tennis and basketball. In partnership with the NBA, they launched the Virtualised Live Match Tracker, using computer vision technology to capture data, which is transformed into 3D visualisations in real-time, offering customisable viewing angles and immersive experiences for fans.
“Table tennis was a good starting point, and we brought a lot of products to market at that time. It was important from a validation standpoint, and it proved a lot of internal hypotheses, making it easier to take the technology to more popular sports such as basketball and tennis. You have to start somewhere, you invest, you see how it works and then you scale it. This is what has happened over the last two years.
“It’s important to note that everything which happens here is also fed into other products. All the streaming products that we offer are interconnected. From where we stand right now, you can see the whole life cycle right from point of collecting the data and steam through to the actual delivery to customers in real-time. Latency is measured in just a couple of seconds, which is impressive. It’s collected here, processed in the cloud, brought back here, and delivered in that time.”
Luka explains how the speed and depth of data enables a micro betting option on the basketball match being played in front of us whereby there is a four-five second window after every basket that punters can place a bet on who will score next. A fast-paced engagement tool, the statistics and micro betting options on offer provide additional value in cementing engagement.
At previous exhibitions. operators would appreciate the technology powering Computer Vision but question how – in numbers and metrics – it benefits the bottom line. Now, the value proposition isn’t in doubt.
“AI applications are nice, but it must create value and generate revenue,” says Luka. “Our customers want to see players engaged more deeply as it helps with customer retention. Operators are seeing their engagement figures growing because users are attracted by the tech and stay longer in the game. The next step for Computer Vision is virtualisation where you can create a completely virtual space, which is what we have for NBA. You can observe a match from any angle, so every user gets a different experience. The possibilities of where we can go with this are endless.”
GEN AI
It’s not just through Computer Vision and 4Sight Streaming that Sportradar is transforming sports consumption. Offering deeper insights and more interactive experiences for fans, broadcasters, and betting operators is the recently introduced “gen AI” suite of products within its ad:s multi-channel marketing service. The tool integrates gen AI with Sportradar’s live data to automate the creation of personalised, real-time audio advertisements for sportsbook and casino operators.
The technology enables operators to efficiently generate creative content during key sporting and gaming moments, scaling their advertising efforts across podcast streaming services and internet radio networks. The automated adverts can dynamically update information such as odds, jackpots, and upcoming events, enhancing brand awareness and customer acquisition.
“The gen AI features are dependent on all this deeper data and tracking. For example, we can automatically create a pre-match or post-match podcast based on the data – even to the point of using a famous person’s voice.
“With the automatic advertising creation, we’re understanding how this deep data can create value for operators looking to offer more media content without dedicating the resources of a big team to it. Through this technology you can efficiently and effectively create such hybrid content.
“With Computer Vision we have reached a level of maturity where now it’s about scaling to the right sports, trying to make sure we have consistent deep data across our portfolio so that our products offer a similar experience. We don’t want the depth of data to be wildly different in scope from one sport to the next.
“The next big thing for us is where gen AI is going; to create a variety of content based on all the data we capture, from probabilities, predictions and enhancing streaming through to editorial content, virtual experiences, podcast creation. Our goal is it make life easier with sophisticated technologies that are accessible to everyone.”
