Cover illustration for How an EU Running Coach Spots Neuromuscular Fatigue in Minutes

What the AI found

“"Across four weeks, your client's cadence dropped 6% only on Thursdays following Wednesday evening sprint sessions, despite normal resting heart rate."”

Before

Three client WhatsApp threads and unread Garmin exports

After

One 4-minute Monday audit that spots fatigue early

The same system, three states — real screens, not a screenshot

1Starting
Apple Notes
  • Mark: complaint of heavy legs on Friday tempo
  • Garmin weekly mileage: 58km (on target)
  • WhatsApp: "Sleep was 8 hrs, HR fine, but strides felt dead"
  • Need to cross-check Thursday split cadences vs baseline
2Working
Claude

Prompt

Analyze this runner's last 4 weeks of Garmin interval data against their post-run notes. Check cadence, stride length, and ground contact time balance against session time-of-day. Flag any mechanical degradation that precedes subjective fatigue reports.

Analyze this runner's last 4 weeks of Garmin interval data against their post-run notes. Check cadence, stride length, and ground contact time balance against session time-of-day. Flag any mechanical degradation that precedes subjective fatigue reports.

AI

Across four weeks, your client's cadence dropped 6% only on Thursdays following Wednesday evening sprint sessions, despite normal resting heart rate. Ground contact time asymmetry shifted 2.1% to the left on those specific morning runs. Suggest moving Wednesday intervals before 18:00 or adding 12 hours of recovery before tempo efforts.
3Implemented
Google Sheets

4m 12s

Audit time / client

1 pattern

Flagged deltas (Wk 4)

0

Unread client exports

PractitionerResources in use

How an EU Running Coach Spots Neuromuscular Fatigue in Minutes

How an independent running coach turned disparate wearable files into a four-minute weekly biomechanics check.

An independent endurance running coach working with 12 amateur marathoners, EU

Tools used

The real tools used here — swap any for your own equivalent. Each links to how we’d set it up.

3 min readWellness & AI editorial
1

Before anything was set up

Sunday night meant twelve open tabs, raw Garmin Connect exports, and mismatched WhatsApp messages from athletes reporting heavy legs despite normal sleep metrics. The coach had columns of pace, cadence, and heart rate across four weeks for an amateur marathoner who was stalling on Thursday workouts. She knew subtle fatigue was accumulating, but tracing step rate regressions across interval sets manually took twenty minutes per runner, so adjustments were made on gut feeling rather than verifiable patterns.

Apple Notes
  • Mark: complaint of heavy legs on Friday tempo
  • Garmin weekly mileage: 58km (on target)
  • WhatsApp: "Sleep was 8 hrs, HR fine, but strides felt dead"
  • Need to cross-check Thursday split cadences vs baseline
2

Resources, doing its job

Using the practitioner movement audit resource, the coach downloaded the raw interval metrics from Garmin Connect and pasted the tabular data directly into Claude alongside the athlete's qualitative log. The prompt instructed the model to cross-examine mechanical markers—specifically cadence and ground contact balance—against the timing of midweek workouts. Within seconds, the model isolated the single training variable that standard readiness scores had missed.

Claude

Prompt

Analyze this runner's last 4 weeks of Garmin interval data against their post-run notes. Check cadence, stride length, and ground contact time balance against session time-of-day. Flag any mechanical degradation that precedes subjective fatigue reports.

Analyze this runner's last 4 weeks of Garmin interval data against their post-run notes. Check cadence, stride length, and ground contact time balance against session time-of-day. Flag any mechanical degradation that precedes subjective fatigue reports.

AI

Across four weeks, your client's cadence dropped 6% only on Thursdays following Wednesday evening sprint sessions, despite normal resting heart rate. Ground contact time asymmetry shifted 2.1% to the left on those specific morning runs. Suggest moving Wednesday intervals before 18:00 or adding 12 hours of recovery before tempo efforts.
3

The finished system, running on its own

The coach now runs the playbook every Monday morning. She exports the prior week's CSV exports into a single raw folder, feeds them through Claude alongside the prompt template, and pastes the resulting two-sentence brief directly into the client's training plan. No manual charting, no cross-checking heart rate against cadence spreadsheets on Sunday nights. The client receives one targeted adjustment before their Tuesday intervals, while the coach keeps her review time under five minutes per runner.

Google Sheets

4m 12s

Audit time / client

1 pattern

Flagged deltas (Wk 4)

0

Unread client exports

22m to 4m

Weekly audit time per client

3 in 4 weeks

Mid-week load adjustments

Garmin ConnectActivity and telemetry capture

Standardized CSV exports for cadence, pace, and split times across all clients.

ClaudeMovement data audit and synthesis

Accurate handling of multi-column athletic data and strict obedience to analytical formatting.

Google SheetsWeekly client overview dashboard

Simple, non-proprietary log that preserves raw values without subscription lock-in.

These are the tools used in this story. Any can be swapped for an equivalent you already trust.

Explore the practitioner movement audit resource

This story runs on Resources. The tools and prompts above are the real build — swap any tool for your own equivalent and follow the same steps.

Suggested for you

Based on what you've been reading — always learning.

See all →