Over the past couple of days, I had the pleasure of attending and speaking at the SportsPro AI conference, held at the iconic London Stadium—previously the venue for the 2012 Olympics and now home to West Ham United. The event attracted over 400+ attendees from across the sports ecosystem, including industry leaders like ourselves at Stats Perform, major tech companies (NVIDIA, IBM, Adobe, OpenAI, Microsoft), leagues/federations and some intriguing start-ups.
The sessions covered a wide array of AI applications in various sports—soccer (both men’s and women’s), NASCAR, sailing, tennis, World Athletics, American football, and the Olympics. Discussions ranged from monetisation strategies and fan engagement to AI-driven commentary and translations, broadcast automation for lower-tier events and broadcast augmentation for higher-tier productions.
Almost every speaker on the day spoke about the value of good storytelling and how AI can help find new stories and new ways to tell those stories.
I was glad to hear that given it’s exactly what OptaAI Studio, Opta Stream, Opta Points and Opta Stories deliver!
Another recurring theme was that AI in sports isn’t entirely new (followers of Stats Perform will know we’ve been applying it for over a decade), but now the focus is on how to harness Generative AI technologies for even greater impact.
This was something I also covered in my keynote presentation on the second morning, where I highlighted the groundbreaking work we’re doing with OptaAI and how we’re leveraging Generative AI models to scale data collection, expand the language of sports, and utilize our content through both textual and visual prompts.
I specifically showcased how we’ve developed the language of sport using our raw data and AI-driven predictions, which provide compelling, interactive content. I then delivered a technical deep dive on how our Opta Vision system uses Generative AI to generate complete tracking data at scale. Additionally, I highlighted our Opta Points product and demonstrated how we can now make player-specific predictions that were previously unattainable. I concluded by explaining how we can “hyper-scale” content generation using AI in two ways: via an “AI-push” (where we automatically push content and insights) or an “AI-pull” (where users can request differentiated content themselves). This represents a seismic shift in how sports content is consumed.
Interesting AI applications were also rife in other sessions, including neat ideas from SailGP, who demonstrated the use of AI for automatic censoring of sailors’, often colourful conversations during live broadcasts, along with remote production techniques aimed at reducing the carbon footprint—a clever and forward-thinking initiative. Another company showcased how AI could be utilized to design and optimize stadium construction, which was equally impressive.
Three key themes stood out to me overall from the conference:
-
AI To Provide More Interactive Content:
Rights holders and broadcasters are at the forefront of applying AI to grow, capture and hold audience attention. They’re enhancing fan interaction through augmented reality, live predictions, and real-time data, whilst also expanding sponsor inventory and value. World Athletics spoke about the power of predictions, explaining that during events, for them, more predictions meant more fan engagement, becoming some of the most viewed content on their site.
We see the same with our Opta Supercomputer predictions. Entire analysis shows are now built around these, and articles featuring Opta Supercomputer win probabilities for major tournaments, or whether a team will secure a promotion or relegation spot or make the finals regularly drive more fan comment and interaction than subjective human predictions.
-
Data is the Gold! Historical Data and Archives are a Gold Mine:
Archival footage and historical data are emerging as powerful storytelling assets for rights holders. Fans are showing a growing interest in comparing current athletes with legends of the past, and AI is helping unlock the potential of vast archives. By creating compelling content that connects past and present, rights holders can attract both nostalgic fans and new audiences alike. They can provide new context and new storytelling angles. Finding the magic in the variety of stories to be told means this can only be achieved with accurate and deep data. AI tools can better leverage these archives, deepening fan engagement and opening up new revenue streams.
An interesting sub-topic of this theme is data’s role in unlocking archives and powering storytelling for past and current athletes directly, as they seek to become legitimate media entities on social channels and beyond (see: Cristiano Ronaldo and Jude Bellingham’s successes on YouTube). Athlete-led storytelling exponentially expands the number of stories to be told and AI is expected to play a starring role in enabling this.
-
GenAI Enables Hyper-Personalization to Heighten Fan Engagement:
The sheer amount of content that can now be created at speed allows for tailored content for individual fans, moving beyond broad, one-size-fits-all tactics delivered days after the event has ended. Rapid and tailored content creation can skyrocket fan engagement. Fans can now receive timely, customized content based on their social media profiles, location, or favorite athletes, making them feel uniquely connected to the sport. It can also provide a much greater sense of “being there” without actually being present.
These three themes align closely with how we view the opportunities of AI, something we’ve been actively working on for a decade at Stats Perform, and with the topics I covered in my own presentation at SportsPro AI, which seemed to be really well-received by the audience and has triggered lots of follow-up conversations.
It was a thought-provoking event, a privilege to represent the work of our teams at Stats Perform on the stage and great to connect with so much optimism about the value AI is already bringing and will continue to unlock for sport.
