The Influence – and Limits – of AI in Sports

We love the humanity of sport. Genuine emotions of winners and losers, bonds among teammates, and a sense of tradition appeal to us. So, too, does sport’s unpredictability, the kind that supplies us upsets, holes-in-one, and unexpected sellouts.

But there are a lot of nuts and bolts that underlie how we get to those glorious moments of human achievement and failure. Human elements factor in, like coaching, athletic trainers, grounds crew, ticket sellers, and even mascots (assuming we’re ok with including anthropomorphic horses, coyotes, and phanatics in a somewhat broad definition of humanity). However, there’s also a lot of, um, math.

“Let's start with the basics. Brandon, do you want to explain? Use your professor voice and explain to us what AI is,” requested host Lauren Allison of Multiplier of one of her panelists at last week’s Global Sports Business Summit presented by Southern Methodist University and the Dallas Sports Commission . 

“Yeah,” replied Brandon Nutting , a vice president for data science and partnerships at MVP . “It's really at scale math.” 

“I could obviously spend a ton of time talking about all the different types of AI,” he continued professorially.” But it's all just math at scale with computers.”

Back in the day, some of us might have used calculators to help us compute batting averages in real time. And the prevalence of analytics in trying to direct on-field results has been prominent in sports at least since the book Moneyball became part of the national consciousness. But this session focused on how front offices use analytics and its buzzy minion, artificial intelligence, to succeed in the sports business outside the playing surfaces. According to Revenue Over And Above Replacement (ROAR) Founder Adam Grossman , it actually starts in the non-virtual world.

“A common question in sports is, why do sports partnerships in the first place? It is the emotionality. It's the human connection with sports,” he said of the sponsorships teams and leagues sell to companies that want to be associated with their brands. “That can seem like a very qualitative thing. Like, how do you know how somebody actually feels about sports? Well, you could actually do it, use a quantitative approach to say, ‘What are they actually saying, particularly saying organically in social media’ and you could actually put a score on that.”

Panelist Katie Morgan works for the Texas Rangers Baseball Club where her title is “Vice President, Business Analytics & Ticket Strategy.” She actually has to utilize scores like the ones Grossman described.

“I think at the root of it, it's made us more efficient. A lot of what we do is using AI at its core. So you think about machine learning, and that's a lot about what Adam just talked about,” she said. “We've been using predictive models for the last 10 years.”

I worked with Katie at the baseball team a little before that 10-year window in which analytics began to become the new normal. When I saw her at the event, I told her I was jealous because I would loved to have had access to the audience data she now commands during my time there. It would have helped me craft more efficient television and radio commercials. She said the club utilizes a platform through Major League Baseball (MLB) that I would have found useful.

“It takes all of the data that we have, all of the data that Major League Baseball has about our consumers, and now we're delivering them the appropriate targeted ad for either season tickets, individual tickets, maybe our brand-new Rangers Sports Network and where to consume that.

“We are putting new tools in the hands of our sales reps,” she said. “It's really allowed both my team, from an analytics perspective, but also our revenue generators, to be more efficient.”

I always felt like an important part of my job creatively was to put consumers in a frame of mind to be responsive when they interacted with a salesperson. I knew that was ultimately how my salary got paid and I wanted all the tools I could get to make it happen effectively and thus keep my job. Grossman knows he has to show his clients how his company’s AI implementation can improve their revenue generation.

“If you're going to make these investments, what's the return on investment? How is it going to drive new revenue? And I think one thing that these tools do allow is not only looking at individual revenue streams, like partnerships, but overlapping multiple different revenue streams to maximize revenue.”

He listed ticket sales, sponsorships, and media consumption as revenue streams they could measure, citing an example of a baseball team’s reworking of home plate signage in which their models predicted future exposure for sponsor messages based on past TV exposure. Nutting explained how digital technology can allow a TV viewer in Canada to see Canadian brands on hockey dasherboards while those in the United States see U.S. advertisers.  

“That requires a ton of artificial intelligence to do. It requires to find those boards and replace them and keep tabs on all the fees that are going out to each individual network. So that's one way that they are using that for a huge impact on revenue stream, to be able to eliminate waste across every channel that they run, and every game has up to five or more feeds that run for every single game that are all separate revenue streams for that individual game.”

So getting sophisticated about improving one’s sports revenue numbers requires a lot of, well, math. Ultimately, though, there’s still human ingenuity both in front of and behind the AI that sports entities are using. Nutting framed the equation with an example from outside the sports world.

“You think about the space race, and everyone likes to joke that we went to the moon with less computing power than we have in our pocket. It's not exactly true. We had a lot of human computers that made that possible, that all had to work behind the scenes, doing very complex calculations. And as computers have evolved and we've been able to do more complex calculations, we've just added more to them.”

Someone had to decide that putting a man on the moon before the end of the 1960s should become a NASA priority and sell that to those who controlled the money to make it happen. Morgan suggested it works the same way in sports.

“You can use AI to make you more efficient. But then the rep comes in and really has that human-to-human interaction to fulfill that sale, right? But for everything that we do internally, whether it has AI or it's just data-driven to some perspective, if we're doing forecasting or dynamic pricing using AI models, there is always the human element of we understand this business to a different degree than the computer does. We have lived it. We have seen fans come in our building,” she said. “Even though every year those decisions become more and more driven by data, you still need to weave a human element into it.”

Immediately after Morgan said the above, the audience saw the human side of one of the AI gurus, as Nutting made a clearly painful admission.

“I can’t believe I'm going to say this: you can't measure everything. Oh, geez, hurt, hurt. There's just things I know,” he said through a grimace. “There's just some things that you cannot stick math on all the way. You can do parts of it, trust me, and people will continue to try, but you can't do all of it in math. And so that's where, like, AI ends, right? It has to be a mathematical representation of something. So you can have stand-in numbers and things that account for things, but it can't account for everything. So that's why the combination (of humans and AI) is super needed, because someone needs to understand the problem that you're throwing at it as well.”

One of the brilliant features of modern AI is that user-friendly interfaces enable a broad cross-section of people, even those not qualified to be NASA engineers, to use big math. But human learning still often outranks machine learning. Allison commented to the audience of current and aspiring sports business professionals that “using ChatGPT is like a really good intern, but they still don't know everything. They haven't been in the business as long as a lot of you have.”

The complex math that is AI serves a basic human desire, one that we all have and that is especially prominent in a sports context: we want to win. A win could come on the field or through the sale of something that enhances the experience of a sponsor or fan. We’re all chasing the human joy that comes with such victories. Analytics in sports, ultimately, is a vehicle for getting us to the humanity.

Rush Olson has spent two-plus decades directing creative efforts for sports teams, broadcasters, and related entities. He currently conceives and executes content projects through his companies, Rush Olson Creative & Sports, FourNine Productions and Mint Farm Films. Through MFF, he’s at work on biographical documentaries about Nancy Lieberman, Sidney Moncrief, Pudge Rodríguez, Ed Belfour, and Bob Lilly as well as documentaries about women's soccer (Raising Her Game) and The College Gridiron Showcase (now streaming on VICTORY+).

Subscribe to @MintFarmFilms on YouTube to see excerpts from the documentaries.