Why I built Endurance Signals

Key points

Endurance Signals started as spreadsheets I built to replace tools I was paying for: metabolic profiling, HRV analysis and plan management.
It turns one club training plan into a plan for each athlete, built around their schedule, biometrics and goals.
Training phases, zones, load and flags are calculated by code. AI writes the explanation and drafts sessions. The coach makes the call.
It is in a closed trial with Triathlon Inverness and the EndureIQ group.

Where it started

I trained as a scientist before I coached. I studied medicine until a back injury at 21 changed course, then did a PhD in biochemistry. I raced as a professional triathlete and held the national Ironman record for more than five years.

As a coach I kept running into the same problem. The athletes who did best were the ones whose training I could adjust to what their body and their week were telling me. That takes time per athlete, and a club head coach does not have it.

About two years ago I started asking a narrow question: where can AI help a coach find the signal in their athletes' data, without taking the decision away from the coach?

Rebuilding the tools I paid for

The first answer was practical. I rebuilt the tools I relied on. Metabolic profiling became FuelingSense, a field-test profiler that estimates VO2max, VLamax and thresholds without a lab. HRV analysis became an engine that reads each athlete against their own normal range. Plan management became a way to take a club block and fit it to one person.

They began as spreadsheets for my own athletes, then web apps I shared with clients. Endurance Signals is where they come together.

What it does for a club

A head coach writes one plan for the club. Each athlete can follow it as written, ask the platform to adjust it to their week, or do their own thing. Race weeks and key club sessions stay fixed.

The coach sees a short list: who is outside their HRV normal range, who flagged pain or illness in a morning check-in, whose sessions are drifting from the plan. That is the list of conversations worth having this week.

Computed in code, explained by AI

The rule I hold to is simple. Anything that becomes a training instruction is calculated, not generated. Training phases come from each athlete's own HRV response to volume and intensity. Zones come from their thresholds. Flags come from fixed rules. A language model then explains those numbers in plain words and drafts sessions from a vetted library.

That keeps the output repeatable and auditable. It also means a coach can disagree with a specific number and see exactly where it came from.

Built with a club

Endurance Signals is being tested with coaches and athletes at Triathlon Inverness, where I am head coach, and with athletes I coach through EndureIQ. The aim is not a tool coaches have to learn to use, but one that learns how they want to work. The time it saves is meant to go back into conversation, poolside and one-to-one.

Common questions

What is Endurance Signals?

Endurance Signals is a coaching platform for endurance clubs and coaches. It turns a club training plan into individual plans and turns athlete data, including HRV, check-ins and workouts, into a short list of coaching decisions.

Does AI write the training plan?

AI drafts sessions from the coach's own library and explains the numbers, but training phases, zones, weekly load and flags are calculated by deterministic code. The coach approves what the athlete receives.

Does it replace TrainingPeaks or intervals.icu?

No. It is a coaching tool that works alongside them. Athletes connect intervals.icu or Garmin, and planned sessions are pushed to their calendar and watch.

Can my club join?

The platform is in a closed trial. Clubs and coaches can request early access.