The NBA Stepped Up Its Social Media Game With AI. So Can You

Last month, in their season-opening showdown with the New York Knicks, reigning NBA champions the Boston Celtics made league history by sinking 29 three-pointers over the course of the night, tying them for the most ever scored in a single game.

It was an athletic feat that only a team of humans, playing at the top of their game, could’ve achieved. But when the Celtics posted a highlight reel on X later that evening counting down each record-setting basket, they got an unlikely assist from artificial intelligence.

A program called AVGen, or “automatic video generator,” had been silently monitoring the live feed of the Celtics-Knicks game and, in real time, cutting each play into a ready-for-social media clip. From blocks to buzzer beaters, AVGen filed them all away, complete with metadata recording who did what, when—and how exciting it was on a scale of one to five stars.

According to WSC Sports, the Tel Aviv-based SaaS company behind AVGen, the Celtics made their X post with the help of AVGen. In fact, the company says that every individual team in the NBA, as well as the league itself, licenses the technology.

WSC’s investors have included the owners of the Cleveland Cavaliers, Philadelphia 76ers, New Jersey Devils and Minnesota Vikings. The company says late NBA commissioner David Stern also had a stake. Per Axios, it raised $100 million in an early 2022 series D.

Highlight reels are one of the main ways that the NBA engages with its massive online fanbase, and automating them is a big time saver for the league, which aims to grow its audience globally. But even for industries that have little in common with pro sports, it’s a window into how the booming AI industry has changed brands’ approaches to marketing and digital content, allowing them to scale up their social media strategy in a way no piecemeal process could.

Jesse Shemen, co-founder and chief executive of Papercup—an AI startup that automatically dubs videos into foreign languages—says AI is being used “to speed up existing processes and simplify more complex tasks” across the media and entertainment sector.

“It’s mainly employed as a tool to help media folks do their existing jobs better and faster,” Shemen explains, pointing to both use-cases around creating content (such as by “lip morphing” the faces of on-screen speakers to make them say something new) and distributing it (especially when it comes to “repetitive or laborious tasks like tagging and archiving footage for a news agency that processes large quantities of video clips”).

For WSC, the focus is on making it easier for teams, leagues and other brands in the sports industry to quickly–even automatically–create engaging digital content. Amir Gelman, WSC’s vice president and general manager of the Americas, says social media managers shouldn’t need to spend hours combing through gameplay footage whenever they want to make a new video for their brand account.

WSC first partnered with the NBA about a decade ago, when the league allowed the startup’s founders to test their tech on footage of that year’s G-League games. The pilot was deemed a success, the firm says, and led to a multi-year contract that began with the NBA’s 2016-17 season.

“Ten years ago it was very expensive and very manual having to create content,” says Bob Carney, the NBA’s senior vice president of digital and social content. “We would have a team of video editors that were working at night, and there was obviously a limit to what that team could produce.”

Bringing AI into the picture, Carney continues, allowed the league to scale its brand strategy globally, routing a now effectively unlimited stream of web content all around the world. The league has even started using AI to translate narrated videos into other languages; more than three quarters of the league’s online fan engagement, he says, now happens outside the U.S.

“There’s so many different touch-points,” Carney says, from social media posts to emails to NBA.com web pages. “If we can localize those and make them feel very authentic to the audience that they’re serving, we think we’re going to see enormous growth.”

AVGen’s automation goes beyond just identifying and labeling gameplay. The software allows licensees to pre-program “if/then” rules that will automatically cut together a video—and even publish it online—whenever certain conditions are met. For instance, the Los Angeles Lakers could tell the system to immediately post footage to TikTok every time LeBron James dunks the ball. Gelman estimates that many of WSC’s partners have close to 1,000 such triggers in place, watching and waiting for their time to shine.

It’s not just basketball, either. WSC’s AI has been trained to recognize key moments in more than 40 different sports, and the company’s client list has included NASCAR, MLS, La Liga, the NHL and the United States Golf Association.

There are also opportunities outside gameplay. The company has been expanding its support for off-the-court data streams such as press conferences, player interviews and sports talk shows. When you add archival footage to the mix, the possibilities seem endless.