Athlete Intelligence

FiyrPod was already capturing GPS load from wearables. Athletes were submitting daily wellness inputs. RPE scores were coming in after sessions. Force plate metrics existed too. The raw signals were there, but there was no layer that translated them into a readiness picture a coach could trust.

In elite programs, a sports scientist handles that synthesis every morning. Most coaching staffs do not have that role. Instead they spend 60 to 90 minutes stitching together disconnected tools, spreadsheets, and intuition before they can even start the session.

Athlete Intelligence became that missing infrastructure: a structured intelligence layer that applied evidence-based scoring across distinct signals and surfaced the output in a workflow a coach could understand in under five minutes.

Every morning was a manual assembly job. Coaches pulled GPS data from one platform, chased wellness updates from athletes, cross-referenced RPE in another tool, and then tried to normalize those signals into something actionable.

That time cost was not a coaching failure. It was the tax of doing manually what a proper system should do automatically. Without a scoring framework, raw numbers were noise. Without a sports science layer, coaches carried the liability of every decision without a way to defend it.

I led research with coaches, worked closely with an embedded performance coach, defined MVP scope, and wrote the product brief engineering used to scope the build. I also designed the full workflow from team dashboard to athlete drill-downs and wrote the system copy across labels, tooltips, and empty states.

The key work on this project was making complex product decisions legible: what the system should measure, how those signals should be weighted, what should be surfaced first, and what needed to stay hidden until the moment a coach needed to explain a decision.

Each system measures one thing, from one data source, and produces one score a coach can act on. Readiness translates six self-reported wellness inputs into a daily readiness score. Pain and Soreness uses body-map selections and severity ratings to surface injury risk. Total Load normalizes GPS metrics against personal baseline. Force Plate flags biomechanical changes. RPE captures perceived exertion and session load.

The complexity is in the model, not in the interface. The team list is ranked by risk rather than name, so the sort order itself becomes the workflow. Detailed views exist for explanation, not for daily browsing.

At 30 days post-launch, coaches reported a 95%+ reduction in time spent on morning synthesis, moving from 60 to 90 minutes down to under 5. Eleven of fourteen coaches stopped maintaining their spreadsheet within the first week. Satisfaction averaged 9 out of 10 across 14 coaches.

The most meaningful behavior change was not adoption of a new screen. Coaches stopped feeling the need to justify their process to themselves. That evidence later became part of enterprise partnership conversations.

I assumed coaches would use the drill-down views more often than they actually do. In practice, most daily sessions never leave the team readiness list. The detailed views matter when a coach needs to explain a decision to an athlete or management, not as a routine destination.

If I revisited version one, I would rebuild onboarding so the first session feels immediately useful. The product should open with pre-populated demo data instead of a blank state that asks coaches to do setup work before value is visible.