Reprioritising a Ligue 1 Squad in Eleven Days

Olympique Lyonnais completed a Stage 1 proof of value in April 2026. It was run on Be-Healthy.AI’s existing models and publicly available data: no integration, no club systems, no access to training or medical records.

It was designed to answer one practical question. Can this tell us which players warrant attention, and how that changes?

What happened

On 18 April 2026, five players sat in the Critical band. Eleven days later, on 29 April, one did.

Four of the five had moved to safer levels. A fifth remained elevated but below Critical. In their place, a player who had been sitting in Medium-Low became the squad’s highest priority.

Nothing about the squad’s headcount had changed. What changed was where the attention should go, and the system said so without being told.

Why that is the result worth reporting

A static risk list is easy to produce and hard to use. Run it once and it tells you who looked exposed on the day you ran it, which is a fact with a short shelf life.

The thing worth having is a list that reorders itself. Over eleven days this one did: it released four players it had flagged, held one under continued watch, and promoted a player who had not previously registered. That is what an attention-ranking layer is supposed to do, and it is only visible if you keep looking.

What this does not show

Four players moved out of the Critical band. It would be easy, and wrong, to present that as a result Be-Healthy.AI produced.

Lyon may well have acted on the initial findings. Be-Healthy.AI had no access to the club’s training plans, medical decisions or workload interventions, and therefore cannot say whether it did, or what difference any action made. The observed reduction is evidence that the system detected changing risk profiles. It is not evidence that anything Lyon did caused them, and it is not offered as such.

The proof of value also did not establish that every elevated-risk player would go on to be injured, that every injury could be predicted, or what happens when these signals are embedded in a club’s normal medical and performance workflow. Those were questions for a longer second stage.

What was demonstrated

Six things, over about six weeks of monitoring:

Squad prioritisation. A ranked view of which players warranted closer review, rather than treating a whole squad as equally exposed.

Longitudinal monitoring. Risk shown as a signal that moves, not a one-off assessment.

Player-relative context. Current risk read against each player’s own recent baseline, so an unusual state is distinguishable from a habitually high one.

Competitive benchmarking. Lyon’s squad-level exposure placed against the rest of Ligue 1, so the number meant something beyond itself. The club sat in the upper third.

Dynamic reprioritisation. The reordering above.

Low-friction adoption. All of it from existing models and public data, before any integration work.

Individual players are not identified. This write-up reports squad-level movement between risk bands, which is what the proof of value was for.