Turning video into evidence instead of opinion

In talent discovery, judgement is invisible but expensive. A player’s future moves through impressions, handoffs, and decisions that depend on context held by individuals who happen to be in the right place. As the pool of players grows, this model stops scaling. The answer is not more highlight reels, but a change in what counts as proof.

Instead of subjective evaluation, performance is structured into a continuous measurement system where video, metrics, and decisions are connected. Computer vision is introduced not as a feature, but as the evaluation layer — reading movement from a single-angle phone recording, validating that the result is real, and producing metrics that stay comparable across players, ages, and countries.

In practice, this means designing the platform around a few core principles:

• measurement from ordinary hardware instead of controlled facilities

• verification and anti-cheat built into the capture itself

• standardized metrics that hold their meaning across regions

• every result shipped with its confidence interval and the clip that proves it
  
  

A key focus is removing dependency on physical presence. Camera recognition and kinetic identification — player and ball tracking, 17-point pose estimation, and field homography that converts pixels into meters — are the infrastructure. The product is the proprietary talent-detection layer above it, which turns raw trajectories into indicators: real acceleration, movement efficiency, technical execution, decision speed, and spatial awareness.

Trust plays a critical role. Because the users are minors, safeguarding is designed in from the start: guardian consent, athlete-owned data, and a firm policy against selling minor data. Verification is not a compliance step; it is the reason a club is willing to act on a player it has never watched in person.

The same verified profile is designed to carry into a second use. Once movement is measured, it can be rendered as a digital twin — a planned vertical, subject to agreements with football-game publishers, that would let a child enter the titles they already play using the same data that proves their ability to a scout.
Over time, discovery shifts from reactive to structured. Clubs spend less time travelling to find players and more time evaluating the ones worth evaluating. As the dataset grows, identification improves instead of degrading. The result is not just software — it is a different way of deciding, where talent is provable, comparable, and no longer dependent on who happens to be watching.