Eager to impress, the young players ran and dribbled between the purple cones distributed over the broken grass. Speed, control and ball control were evaluated carefully — but not by a veteran scout.
Not even by a human being.
Instead, Brazilian athletes were being analyzed by an artificial intelligence cell phone application, part of a new wave of tools that promises to revolutionize the way talent is discovered in the country of football.
“We are talking about millions of boys and girls who are never seen,” said Roger Wittmann, a German sports agent who created Cuju, one of the observation apps that has been gaining ground in Brazil. “This is a great chance for them to show up.”
Platforms like Cuju quickly attracted hundreds of thousands of users in a country where playing football professionally is a dream for many people. The tools have also caught the attention of large clubs, some of which are already using apps in recruitment.
Brazil, where football is deeply linked to everyday life, exports more talent to the sport than any other country in the world, with some athletes earning millions of dollars at major European clubs.
Continues after advertising
AI scouting platforms are already common in Europe, where football talent assessment has long been based on metrics and statistics. However, in Brazil, deep economic and regional inequalities have historically made the standardization of this process difficult.
Instead, the discovery of talent in Brazilian football was left to the country’s legendary scouts. Known precisely as “scouts”, these veteran talent hunters usually spend decades scouring amateur matches, neighborhood championships and school tournaments, from the Amazon Rainforest to the backlands, looking for the next big star.
But now, AI technology can identify promising talent faster — and, in some cases, better — than the human eye alone. It can also reach more athletes in a territory as vast as Brazil, giving those who live in remote regions, where few scouts venture, a chance to be discovered.
Continues after advertising
Most AI observation platforms work by analyzing user-uploaded videos or recorded workouts directly in the app. Considering a wide variety of skills, from speed to ball control, these tools generate a score and add athletes to a database. There, human agents looking for talent can find them — or the applications themselves can introduce them directly to specific clubs.
For now, at least some recruiting still remains in human hands. But the advancement of AI is already fueling discussions about how much of this work should be handed over to machines.
On a cold Sunday morning, a few dozen teenagers lined the worn concrete bleachers of a modest stadium in the interior of São Paulo, clutching worn and muddy soccer cleats.
Continues after advertising
Many had come from cities hundreds of kilometers away to participate in a trial that could earn them a place on a local team in Aguaí, a quiet town of 30,000 inhabitants. The athletes, all aged between 14 and 19, had been selected based on scores obtained from an artificial intelligence application. Now, they hoped to show what they knew how to do live.
On the side of the field, some boys did one-minute exercises while the app captured and evaluated their movements in real time. Then, on the pitch, the teenagers competed for the ball under the watchful eye of a team of coaches.
Davi Barossi, 18, soon attracted attention. He dribbled past two defenders and sent the ball into the corner of the goal. Barossi had traveled 10 hours by car from Santa Catarina. “I’m here chasing my dream,” he said, the day after watching the Brazilian team compete in the World Cup.
Continues after advertising
Nathan Moraes, also 18 years old, from Pará, had more difficulty. After an unsuccessful cart, he left the field limping, with an expression of pain. “Every opportunity that comes your way, you have to give it your all,” he said, massaging his locked shin.
During a break, the players teased each other, devoured chopped fruit and drank water without stopping. “I’m in second place on the app”, boasted Moraes. “And you, what is your position?”
On an already pretty beat-up cell phone, Barossi showed his own metrics. He said he worries about being shorter than many players his age. But, after exercising through the app every day, she had risen to the top 30 nationally in her age group. “I’m always kicking the ball and recording,” he said.
While most experts agree that AI tools can help define more accurate and standardized criteria for evaluating players, they warn that the technology also has blind spots.
Metrics may favor taller or stronger athletes, missing out on less conventional talent. Although much of Brazil is connected, these applications remain less accessible for poor athletes without good internet or a quality cell phone camera. Furthermore, users can delete or change uploaded videos as many times as they want, increasing their score without this necessarily reflecting their real capacity.
There are also those who believe that, no matter how advanced it is, artificial intelligence simply cannot surpass the trained eye of a professional scout.
“It’s a gift that God gives,” said João Maradona, a Brazilian scout whose work in the remote Northeast revealed several athletes who later made the Brazilian national team. “No one can teach you to see, in just 15 or 20 minutes, that raw talent that is really special.”
It’s no surprise that those developing AI observation applications have a different view. In a small startup office in São Paulo, on a recent afternoon, analysts working for the Footbao app analyzed videos uploaded by athletes. With quick mouse clicks, they assigned ratings to each player in about two dozen categories.
The grainy videos had been selected by artificial intelligence as the most promising among tens of thousands sent by players from all over Brazil. After the human team reviewed this material, the AI ranked the athletes using a formula and generated a detailed report for the clubs.
Artificial intelligence was still being trained. But the goal was, in the future, to automate the assessment, reducing human errors and making the observation process more data-based.
“We’re not trying to take the scout’s job away,” said Nick Rappolt, the startup’s chief operating officer. “We are making observation more efficient and more economical.”
In many ways, advanced technology is already transforming football, as major clubs test AI to analyze matches, prevent injuries and create game strategies. For technology advocates, recruiting is just the next frontier.
The growing role of technology was evident on a recent morning at the Santos FC training center. The base team ran from one side to the other on the field while a drone flew over the place. The recorded images could be analyzed by AI and used to measure players’ performance.
Santos, where legends like Pelé and Neymar Jr. began their careers, entered into a partnership with Footbao, with the aim of using the application in recruitment and staying ahead of rival clubs.
“We can’t be everywhere at the same time,” said Carlos Antônio Anunciação, the club’s athlete recruitment coordinator. “Today, with the help of technology, we can go much further.”
Still, he was preparing to travel more than 1,000 kilometers the next day just to see a promising player up close. The information, according to him, came via a WhatsApp message sent by a veteran scout. “I can’t resist seeing it live,” he added.
Back at the sieve in the interior of São Paulo, the last match ended just before noon. Sweaty and exhausted, the teenagers leaned against a rusty fence, waiting for the results to find out if they had been approved to join the local team.
Barossi and Moraes were among those chosen, along with six other players selected for the squad. The rest, with their shoulders slumped, began to disperse.
The new recruits, smiling from ear to ear, headed to the locker room. “This is the opportunity I was looking for,” said Barossi.
This article was originally published on The New York Times.
