The New York Jets have been pretty uniformly terrible on the field under the ownership of Woody Johnson, but the organization has lately been recognized as an innovative and industry-leading presence in integrating video game rankings and loutish patrician antics into its front-office processes. That hybrid approach has admittedly not yet borne fruit in terms of top-line indicators like wins, or recording even one interception over a 17-game season, but this sort of work takes time and evolves alongside the technology available. A team must accumulate some more abstract wins off the field before those successes show up in the standings.
This work is not glamorous or public, but as New York Jets Chief Analytics and Data Officer Iwao Fusillo told Sports Business Journal, he's already seeing results despite just joining the organization in January after a long career in corporate America. The Jets organization, he said, has already made strides, because "91 percent of the Jets’ front office now uses Microsoft Copilot on a day-to-day basis—up from 'a handful' about a hundred days ago—with users averaging two-to-three prompts per day."
How should Jets fans feel about the fact that a large and suspiciously precise percentage of the front office is now firing questions into the widely loathed and overwhelmingly unreliable AI modality in Microsoft's office technology? My friend, I am not in the business of telling Jets fans how to feel about things, and as a rule seek to limit engagement in that area to the extent possible. But it seems clear that Fusillo is pleased with the adoption of this bummy technology so far. "I call that level one, or horizon one," Fusillo told SBJ, "which is adoption." (The additional two levels, or horizons, have to do with "deeper levels of workflow automation," SBJ's Rob Schaefer explained.)
"Do we have large business gains from that level one?" Fusillo continued. "Not really. But have we changed the culture of the entire front office? Yes. To think AI-first."
It's very easy to goof on the Jets for patting themselves on the back for getting 91 percent of their front-office workers to avail themselves of one specific universally reviled AI function, whose own terms of service include the words "Don’t rely on Copilot for important advice." But in the interest of fairness, it's worth noting that AI-adjacent technologies like machine learning have long been a big part of teams with smarter front offices. Forward-thinking organizations in every sport have been using machine learning to solve discrete problems and unearth precious fragments of edge from dauntingly large data sets; commercial AI, of the kind that shows up in chatbots and web search or email, differs mostly in terms of its broader remit and interface.
A sports analytics department would use machine learning to organize and analyze the vast amounts of data that sports run on; the good organizations are the ones that make more intelligent use of it. In those front offices, this burgeoning computational capacity is brought to bear on longstanding questions in the hope of generating ever more refined answers. The effectiveness of the technology, in this context, is limited mostly by the capacities of the people using it, who must know how to ask the right questions and parse the answers, and also enough about how it all works to be able to refine that process where it needs refining. A laptop can be used as a hammer in a pinch, but that is not how it's most useful. It depends who's holding it, and what they know how to do with it.
Fusillo is right that the less obviously useful engagements with this kind of work at level one, or horizon one, are necessary in getting skeptical or behind-the-game front-office workers on board with a technology sold to their bosses in breathless and bloodthirsty terms as a way to automate those very workers out of their jobs. He noted that the Jets already have "over 20 deployments" in level two, or horizon two, which includes "areas like sponsor prospecting research, revenue reconciliation and, on the football side, structuring data from physician evaluations of players at the draft combine, among other use cases."
A complicating factor here is that these are the New York Jets, which means that it is difficult not to read them glorying in having done the simplest and least useful bit of AI integration imaginable as a boast about this expensive and fragile hammer they've been getting into as of late. Because it is the Jets, it is much easier to see this as a mandate from the posh dingus in the ownership suite: that everyone working for him should use a technology that said dingus doesn't understand at all, but which he has heard is very urgently The Present and also, in some grandiose and vague ways, The Future. Even if AI and machine learning technology could be useful for sports front offices, that does not mean every team will successfully embrace or even meaningfully comprehend what those uses might be.
It certainly does not mean that making more use of this technology—for instance, asking Copilot to help out with your email inbox and then sighing heavily when it haplessly cocks up that task in potentially legally actionable ways—is making better use of it. The former is much easier to accomplish than the latter, and to a Woody Johnson type, the two would feel more or less the same. If this sort of computational capacity is finally a tool, which for all the messianic pomp attending AI is still about the nicest thing that can be said for it, then there would necessarily be more and less skillful ways to use it. A technology like that could much more easily be used in a proudly backward workplace to create another useless annoyance for workers.
A technology like Copilot, which delivers answers to queries not wisely but too well, would produce outputs just as reliably as the more complex machine learning systems that would have to be maintained and operated by more highly skilled workers, although those outputs would be more opaque and low-quality. They would be very hard to trust or verify. But to a boss without the capacity or attention span to discern between mere activity and actually useful work, or between An Answer and meaningfully useful intelligence, it would all look about the same. For an aging princeling without much interest in this or any other kind of work, who mostly understands his job as performing a series of gestures and noises that indicate being in charge, both would be equally useful.
Such a boss would not be able to grasp the analytical utility of this kind of technology, but would be reflexively inclined to credit it more than the work done by the actual people he pays. What a boss like that would understand about a technology like this would be limited by the laziness and anti-human ideology that defines his social cohort; it would be less about what that technology might do than the mere act of doing it. Such an owner would understand and appreciate AI primarily as a thing that generates an output, and that might someday generate that output where a person, with all their needs, smells, and rights, had previously been required. The process and the outcome would ultimately be immaterial to this kind of boss; the doing of it would be the thing. A boss like this would very much want to be "AI-first," but would not really care about what that means or does. Something like an announcement that his NFL team has increased AI usage, by 91 percent, would probably do the trick for him. There are levels and horizons to all of this.
