Wimbledon’s new Catch Me Up AI feature promises to keep fans up to date at the tournament – after it irons out some of the wrinkles
The latest feature to come out of IBM’s partnership with Wimbledon will keep fans engaged from the early stages right through to the final with dynamic player insights
The last five years of IBM’s long-standing partnership with Wimbledon has seen AI take ‘Centre Court’ as the latest technology the tournament can leverage to drive fan engagement and enhance customer experiences.
IBM
has been the official technology partner for the Championships for over 30 years since the initial partnership was signed in 1990, and this year tournament organizers, the All England Lawn Tennis Club (AELTC), launched its latest generative AI feature, Catch Me Up.
Unveiled on 17 June 2024, Catch Me Up promises to offer a richer and more engaging fan experience throughout the tournament. The tool provides AI-generated content for player cards with recent stats and play styles, which are continually updated throughout the tournament.
These player cards will be personalized according to user preferences using location data and their myWimbledon profile, and will feature both pre-match and post-match content.
The pre-match content will include information on the player’s career and their recent form, as well as win probabilities, while the post-match content will feature key statistics from the contest and a highlights package to catch up on the match.
Catch Me Up is powered by Granite, IBM’s flagship LLM foundation model available through its Watsonx AI platform.
This foundation model is then trained on years of structured data collected during previous tournaments and unstructured statistics from external sources to provide accurate content about the player’s recent form or career history.
Get the ITPro. daily newsletter
Receive our latest news, industry updates, featured resources and more. Sign up today to receive our FREE report on AI cyber crime & security - newly updated for 2024.
Chris Clements, digital products lead at the AELTC explained the organization centers fan experience when developing their AI implementation strategy, and Catch Me Up will help fans stay engaged throughout the tournament.
“We are committed to offering fans the highest quality Wimbledon experience, whether it's in person or digitally,” Clements stated.
“Generative AI allows us to scale our ability to provide different types of content for fans wherever they are in the world in a way that's personalized for them. This year's new Catch Me Up feature will make it easier for fans to follow the key storylines as they emerge throughout The Championships."
Catch Me Up integrates last year’s AI-generated commentary
Each Spring, IBM and the AELTC coordinate on how the user experience at the tournament could be enhanced, looking back on what worked and what didn’t from last year’s event.
This year-round innovation workstream begins with identifying the fan personas that they are trying to cater to, with some fans preferring in-depth statistics, while others favor video content or ‘snackable’ content distributed through social media.
These personas are then fed into an ideation workshop each Spring where IBM and the AELTC discuss current trends in sport and technology as well as the two organizations’ priorities to begin designing the features themselves that can realize these goals.
Each Autumn the two come together once again to decide which of these features will be implemented from those ideas in the following year’s championships.
Last year, Wimbledon introduced a generative AI commentary feature that used AI-generated audio commentary and captions for post-match video content.
The feature drew criticism from pundits expressing concerns about it being used to replace human commentators, and has reportedly been scrapped for this year's tournament. However,according to Kevin Farrar, head of sports partnerships at IBM UK, it has been integrated into the Catch Me Up tool.
Speaking to ITPro, Farrar said the feature would not impact its commentary teams and is aimed at improving the AELTC’s ability to provide expanded coverage at the Championships.
“What we wanted to do and what Wimbledon wanted to do was to broaden the aperture, to tell more stories about more matches, and pre-match as well as post-match. So that was the goal and that ended up being rendered as Catch Me Up.”
He added that tennis is in a transitional stage at the moment, with a flood of new talent taking up the reigns of previous legends like Nadal, Djokovic, and Willams, and these features will help introduce fans to the new stars as well as other parts of the game that previously have not enjoyed the same attention.
“It’s about telling more stories that aren’t being told at the moment potentially. You’ve got a lot of new players up-and-coming, a lot of the old guard coming to the end of their careers, how can we tell stories about those other players?” he said.
“These features are enabling that, they’re trying to get stories out there about some of the new players that aren’t being written about at the moment by the content team or journalists - it’s filling a gap.”
Catch Me Up’s debut experiences a few teething problems
Catch Me Up has exhibited some growing pains since it was rolled out at the Championships on 1 July, with some fans pointing out on X that it had made a series of errors in various player cards.
For example, former US Open Champion Emma Radacanu’s profile erroneously listed her as the current British number 1 and having racked up 11 wins this year, despite being ranked British number 3 and winning 14 matches.
All AI-generated content undergoes human checks, according to IBM, so it appears these were missed by Wimbledon’s content team, but as users on X noted, the errors were quickly updated in the Wimbledon app.
Similarly, despite the model being trained on Wimbledon’s house style, fans noted a series of US spellings in the copy generated by the Catch Me Up engine.
Farrar said this was to be expected at this early stage of implementing generative AI solutions, but stressed the importance of recognizing these shortcomings and dealing with them quickly.
“It is a process of continuous improvement. So the model is trained and then there are checks and balances in there to make sure that we can minimize those errors,” he explained.
“That’s the nature of AI. What’s important is that we have ways to recognize those and deal with them quickly, and that’s what we do as you saw.”
He added that as the tournament progresses, both the model and the human teams working on the Catch Me Up content will improve, and the frequency of mistakes should decrease.
“As we go through the tournament that will get refined and improved. The team’s learning as well as the AI model is. It’s a pilot and it's working together to try new things, and each year we try to push the boundaries and get that balance right between tradition and heritage and tech and innovation.”
Solomon Klappholz is a Staff Writer at ITPro. He has experience writing about the technologies that facilitate industrial manufacturing which led to him developing a particular interest in IT regulation, industrial infrastructure applications, and machine learning.