The old way vs the new way
Traditionally, scouting in cycling was a manual process, limited by geography, staffing, and opportunity. Sports directors would attend a handful of domestic races, observing riders firsthand, and making recommendations based on their performances. This method, while effective, is inherently restrictive. Only a small number of athletes could be evaluated, and many international riders often slipped through the cracks.
Enter AI-driven scouting. Platforms like TrainingPeaks revolutionised the way athletes share their performance data. By voluntarily uploading their training files, riders unlocked new possibilities for talent identification. A cloud-based system can now standardise data from all athletes, compute performance scores tailored to different racing roles, and update these scores with each new training file. Suddenly, hundreds of athletes could be ranked in real time, regardless of their location.
How the AI system works
The process begins with data standardisation. Every athlete’s performance metrics (power output, heart rate, etc.) are collected and normalised. This allows for a fair comparison between riders, regardless of the equipment or conditions they train under.
Next, the system computes performance scores. These scores are tailored to the specific demands of different racing roles. A climber’s data is evaluated differently from that of a sprinter. This ensures that the AI can identify talent that aligns with the unique needs of a team. And as athletes continue to train and upload new files, the system updates their progression.
World Tour teams leading the charge
The adoption of AI in cycling is not just theoretical – it’s already happening. GreenEDGE Cycling, the organisation behind the Jayco-AlUla men’s team and Liv-AlUla-Jayco women’s team, recently partnered with ai.io, a UK-based AI technology company. Their collaboration focuses on performance development, talent identification, and even fan engagement. Tools like ai.io’s 3DAT motion capture system and aiScout talent platform are being used to analyse rider technique, prevent injuries, and uncover hidden talent.
This trend extends beyond GreenEDGE. The Ineos team, now sponsored by Netcompany, has made AI a cornerstone of their strategy, aiming to win the Tour de France again within five years. Similarly, Visma-Lease a Bike has partnered with Mistral AI, while UAE Team Emirates-XRG joined forces with Analog. These collaborations highlight a broader shift: AI is no longer a futuristic concept but a practical tool reshaping cycling at every level.
The human-AI collaboration
The one thing that AI can’t replace is human judgment. It acts as a force multiplier, surfacing athletes whose performances might otherwise go unnoticed. Coaches and performance staff still make the final decisions, but the model allows them to see and evaluate more riders, ensuring that no potential star is overlooked.
This collaboration between human expertise and AI is where the magic happens. The algorithm can sift through vast amounts of data, identifying patterns and potential that might escape the human eye. But it is the coaches and staff who interpret these findings, applying their experience and intuition to make the final call.
The lesson
The lesson from this fusion of technology and tradition is clear: AI works best when it combines high-quality data with human interpretation. It is not a replacement for the nuanced understanding of a seasoned coach, but rather a tool to enhance their capabilities. It’s very likely, if not certain, that future Tour de France champions will be discovered by this partnership. The old way simply can’t keep up.