The 2026 World Cup was supposed to be remembered for Spain lifting the trophy. Instead, plenty of fans will remember it for a flag that never went up, a goal that got wiped off the board, and a sensor buried inside a soccer ball that saw something no camera or referee could. The tournament had the most advanced officiating technology in the sport’s history, and by the time the final whistle blew, a lot of people trusted it less than they did before the first match kicked off.
That is a strange outcome for a system built to remove doubt. And it is exactly why a soccer story belongs on a technology site. What happened with VAR and semi-automated offside technology this summer is, almost beat for beat, the same story enterprise technology leaders have been living through with AI for the past three years.
Two Calls That Broke the Internet
Start with Germany against Paraguay in the round of 16. Jonathan Tah forced the ball home in extra time, Germany thought they had a winner, and then VAR pulled the goal back over a foul in the build-up that most people watching at home could barely see. Germany went to penalties and lost, ending the run of a team many had picked to go deep into the tournament.
Then came Portugal against Croatia. Croatia’s Josko Gvardiol buried a stoppage-time equalizer that looked, on every broadcast angle available, completely clean. VAR disagreed. Officials went not to the replay but to the motion sensor tucked inside the adidas Trionda match ball, which had picked up a graze off a Croatian substitute in the build-up, a touch invisible to every camera in the stadium. That touch put a Croatian player offside a split second before the move continued, and the goal was erased.
Neither decision was wrong, technically speaking. Both were exactly what the technology was designed to catch. And both left players, coaches, and millions of viewers convinced the game had been decided by a machine nobody fully understood rather than by what happened on the pitch.
A System That Worked Exactly as Designed
It helps to know what is actually running underneath these calls. FIFA’s semi-automated offside setup for 2026 combines a network of calibrated stadium cameras with skeletal tracking that builds a 3D avatar of every player on the field, accurate to roughly 10 centimeters. Sitting inside the match ball itself is an inertial measurement unit sampling movement 500 times a second, which is how officials can pinpoint the exact instant a player touches the ball, down to a graze too faint for any lens to catch.
The 2026 tournament added a further wrinkle: when a clear violation is detected, the system now sends an automated alert straight into the earpieces of the assistant referees, so a flag can go up almost the moment the infraction happens rather than after a lengthy video review. None of this replaces human officials. The technology hands them evidence and a recommendation. A person still makes the final call, and a digital reconstruction is generated afterward to explain exactly what the system saw.
Judged purely as an engineering achievement, this is remarkable. Judged as a system that needed to survive contact with 90 minutes of live sport, millions of emotionally invested fans, and a sport whose entire appeal rests on trusting what your own eyes just told you, it was something closer to a crisis.
Where the Real Failure Happened
Managers didn’t hold back. England’s Thomas Tuchel and Egypt’s Hossam Hassan were among several coaches who criticized the standard of officiating publicly during the tournament, and the broader complaint wasn’t really about accuracy. It was about a sport that suddenly felt like it was being adjudicated by data nobody in the stadium could see or verify for themselves.
That is the actual failure here, and it has nothing to do with the sensors. The offside detection worked. What failed was everything around it: how the technology’s role was communicated, how far its authority was allowed to creep beyond what fans had bought into, and how little anyone did to manage the gap between what a machine can prove and what a human is willing to accept. You can build a flawless measurement system and still lose the room entirely if people feel the decision was made to them rather than explained to them.
The Enterprise AI Playbook Hiding in a Soccer Stadium
Swap “offside call” for “loan denial,” “performance review,” or “flagged transaction,” and this stops being a story about sports at all. It becomes a near-perfect preview of what is already happening inside companies rolling out AI at scale.
The numbers back that up. Kyndryl’s 2026 People Readiness Report found that 57 percent of organizations have deployed AI broadly or built it into core business processes, yet only 25 percent say they completely trust AI systems operating without a human checking the work. At the same time, 81 percent expect AI agents to make consequential decisions within the next year regardless. Gartner, separately, projects that 55 percent of large companies will lean on automated decision systems for mission-critical workflows by the end of 2026. Adoption is racing ahead of the confidence needed to actually live with it, which is precisely the gap that turned two correct offside calls into a full-blown controversy.
It’s a gap we’ve tracked closely on this site. Our reporting on how confidence in scaling AI has been sliding for three straight years found the same root cause driving both the World Cup backlash and enterprise hesitation: it is never really the model that breaks trust. It’s the absence of a clear, communicated boundary around what the model is allowed to decide, and who is accountable when it decides something people don’t like.
Precision Isn’t the Same Thing as Trust
This is the part of the AI conversation that rarely makes it into a product demo. A model can be accurate down to the centimeter, or the decimal point, and still get rejected wholesale by the people it’s supposed to serve, because accuracy was never the thing standing between a pilot project and real adoption. Trust was.
That’s a big part of why roughly 90 percent of AI projects still fail to deliver real value, according to research tracked by Gartner. The failure point is rarely the algorithm underneath. It’s a rollout that skipped the unglamorous work of defining boundaries, building escalation paths for when the system gets overruled, and giving the humans on the receiving end a real say in how much authority they were signing up to hand over. FIFA had all the technical rigor in the world and still ran headfirst into that exact same gap, live, in front of a global audience.
What Companies Can Actually Learn From a Flag on the Field
A few lessons travel cleanly from the pitch to the boardroom. First, communicate the boundary before you deploy, not after the first controversial call. Fans had years to get used to VAR checking clear and obvious errors; they were far less prepared for a hidden sensor overturning a goal on a graze nobody could see, because that expansion of scope was never properly explained to the people watching. Employees and customers deserve the same heads-up before an AI system’s authority quietly grows.
Second, invest as much in explaining decisions after the fact as you do in making them correctly in the first place. FIFA’s digital reconstructions, showing exactly where every player stood at the moment of contact, are a genuinely good idea poorly deployed too late in the broadcast to calm the room before the anger set in. Enterprises rolling out automated decisions face the same clock: an explanation that arrives after someone has already decided the system is unfair rarely repairs the damage.
Third, treat trust as a maintenance cost, not a one-time launch item. Our look at why small businesses may see AI’s real productivity payoff before larger companies do found that smaller organizations often move faster precisely because they can keep a human closely tied to every automated decision as they scale, rather than bolting on governance after the system is already running at full speed across a global workforce.
The Bottom Line
Nobody is going to rip the sensors out of the World Cup ball, and no company serious about AI is going to walk back its investment because a rollout got messy. That was never really the choice on the table, in soccer or in business. The choice is whether the humans in the loop, the fans in the stands or the employees at their desks, get treated as people who deserve a clear explanation of what changed and why, or as an afterthought to be managed once the complaints start rolling in. FIFA had the better technology and still lost the room. Any company deploying AI this year is one badly explained decision away from the exact same outcome.
