Luka Pataky, Senior Vice-President of Automated Content at Sportradar, details just some of the early stages of AI’s move into sports betting streaming.
Artificial Intelligence is the topic of the day.
It seems like everyone in sport and across all industries is proclaiming their use of AI and the magic it can deliver. The term has become a buzzword and is being thrown everywhere, which does little to help professionals or consumers alike truly understand its real-life applications and benefits.
Put simply, AI brings sport data to life, streamlines previously labour intensive processes, and completes tasks beyond the capabilities of a human. When harnessed effectively, AI is revolutionising the industry.
This has been our focus at Sportradar – pioneering AI to deliver tangible benefits for rights-holders, betting operators, media companies, and fans alike. This means developing models that capture previously untapped datasets at newfound levels of speed and accuracy and transforming this information into next generation products that deliver more immersive, hyper-personalised experiences for fans.
This real-life application of AI was demonstrated at the recent ICE 2025 conference in Barcelona, where we had the opportunity to exhibit how it’s transforming the betting value chain, via a live 3×3 basketball showcase.
Through a single camera, we have trained a subset of AI called Computer Vision (CV) to automate the collection of previously inaccessible sport data, faster than ever before. Whilst the speed and depth of this data capture is significant, it is the AI-processing of these raw coordinates that unlocks the full potential of this resource. If data is the fuel, AI is the engine.
HAVE YOU ALSO READ?: Sportradar releases annual integrity report
Algorithms can be created that recognise and extract meaningful insights from specific sports. The 2023 edition of ICE saw our inaugural exhibition of Computer Vision, showcasing models trained to understand the rules of table tennis, to fuel betting and media opportunities, and score games quicker than a human umpire.
With a larger playing surface, more movement on and off the ball, and a greater number of players and variables, the extension of these models to basketball and team sports is significantly more complex than the relatively confined environment of table tennis. However, AI is capable of being retrained to detect ball trajectory, lines on the court, and differentiate between players, to track the action in real-time.
Most importantly, for a sport as popular as basketball and for a global rights-holder like the NBA, AI-driven contextual insights fuel downstream betting and media markets and open a world of new possibilities for partners and fans. ICE attendees were able to engage with live examples and an ‘under the hood’ view of AI’s capabilities. These value-adding products included our Sportradar 4Sight streaming visualisations, virtualised Live Match Trackers, as well as a micro betting market, which was replicated through a live mobile game relating to the on-court action that spectators could play.
This sequential application of AI is transformative and benefits the entire sports ecosystem. It allows rights-holders to realise the full commercial potential of their data by producing greater and more engaging content for their betting and media partners.
By deriving greater meaning from data, sportsbooks can offer a higher number of betting markets, more accurately predict the likelihood of an event happening, engineer sharper prices and manage their liability. AI also allows the creation of micro markets – live bets on short-term game outcomes, such as the likelihood of a player making their next free throw.
Equally, broadcasts and live streams are being elevated through cutting edge AI-generated virtualisations, perspectives, and graphic overlays. Millions of viewers worldwide can engage with never-before-seen on-screen stats, instantly contextualised to provide deeper understanding and appreciation of the action, plus greater storytelling opportunities for commentary. For basketball, this could be a player’s three-point percentage in a certain quarter, or their steal record across different parts of the court.
The deeper data that can be collected through CV is also key to unlocking many innovative sports applications that utilise Generative AI. Automated commentaries, podcasts, and sports highlights can all be created by leveraging the deeper insights generated by Computer Vision and AI. This presents even greater opportunities within media and broadcasting, as well as for sport betting operators looking to offer more immersive content for their customers.
And not only is AI producing this next generation of hyper-personalised betting and media experiences for fans, but it is also doing so at an extraordinary rate. At Sportradar, we have witnessed this evolution firsthand with our partners. In less than two years since the launch of Computer Vision within table tennis, the technology has expanded to tennis through our partnership with Tennis Data Innovations and the ATP, and basketball via the NBA.
Its application to football, the world’s most popular broadcast and betting sport, is therefore becoming more of a reality. Whilst this presents additional challenges for AI modelling with even more players, a larger playing space and potential weather conditions, to name a few, we’re confident that dynamic animations and live match trackers may be coming to the pitch sooner than we think.
And as sport becomes even more accessible and fan appetite for real-time data and hyper-personalised experiences grows, the reliance on AI to collect deeper and greater volumes of data to develop even more innovative end-user experiences is only set to increase.