The NFL and Microsoft have been sharing a technology playbook for 12 years. But the latest drive will feature more generative artificial intelligence.
On Wednesday, the pair unveiled a multiyear partnership extension that includes the deployment of 2,500 Microsoft Surface Copilot+ personal computers, devices with built-in AI features that will be available to the league’s 32 clubs, roughly 1,800 players, and more than 1,000 coaches and staff. Known as the NFL’s sideline viewing system, these tools are used to access game day images and data, both on the sideline and in the coaches’ booth.
What makes the AI PC unique is that it features a specialized processor called a neural processing unit, or NPU, which allows these devices to run AI workloads locally. Vendors tout these devices for their ability to run AI workloads more efficiently than if data was sent to the cloud or a mainframe server.
The NFL says a new feature on these devices, built with GitHub Copilot, allows users to pull critical information about the state of play on the field and use data to respond in close to real-time to craft the optimal attack or defensive tactic.
“There’s literally seconds in order to process information in between plays or in the coaches’ booth, so that the speed at which the device can do the AI processing is vital,” says Aaron Amendolia, the NFL’s deputy chief information officer. “We can’t go back and forth to the cloud. We can’t wait 30 seconds for your prompt to return a result.”
The prior generation of tablets used on the field allowed players and coaches to access offensive and defensive plays, as well as various camera angles to ascertain how players lined up in the field. In the booth, even with Surface devices, math was frequently done using a pen and paper. But with AI PCs, the NFL expects to automate the data that’s flowing to devices on the field, removing some of the manual toggling that was required previously, while also combining Excel and AI to perform calculations in the booth.
Technology has been a vital team player improving both the viewing experience for fans and play on the field. Throughout the course of the league’s century-long history, key milestones have included the first televised game in 1939, an instant replay system added in 1986, and more recently, Amazon Web Services’ analytics and machine learning deployed to better understand where players are more likely to get injured and react by designing new rules that can mitigate risk.
Amendolia says the league coaches teams on what solutions are best, but ultimately empowers clubs to make the most of their own IT purchasing decisions. The NFL is far more assertive in setting governance policies around key technology pillars like cybersecurity, data standards, and privacy.
There are some exceptions where the league will issue a mandate. The Surface tablets, as an example, are required and were first deployed over a decade ago to improve on-field communication after the NFL and Microsoft struck an initial $400 million deal in 2013. That investment allowed the league to finally retire printers and paper that were used to keep track of plays.
“The clubs are not choosing that technology,” says Amendolia. “They’re giving input into how that ultimate experience is.”
The latest Microsoft hardware integrates well into the NFL’s ecosystem that includes the productivity tools suite Microsoft 365, Teams, and a 5G private network built with Verizon. Sideline tablets are also powered by Azure’s cloud, though the NFL has embraced a multi-cloud approach and works with a variety of vendors, including AWS, depending on what suits each use case.
Amendolia says generative AI is a new layer of technology that builds upon the league’s past investments in machine learning. One use case that the NFL has embraced is a multimodal generative AI search tool that allows the sponsorship department to comb through mounds of data and pull up the exact assets they may need. Previously, this system relied on meta-tagging images, in which key words and descriptions were manually added to each image—a system that was not always reliable during a search query.
Externally, one application of generative AI debuted this year on OnePass, the league’s mobile app that manages tickets and shares event information with fans. Amendolia says that the NFL has made responses more conversational, answering queries with natural language when a fan wants to know the best place to get food in the stadium.
“The idea is that our fans are going to start to have their experience personalized through AI,” says Amendolia.
John Kell
Send thoughts or suggestions to CIO Intelligence here.
