Review your race results with Claude to set better training paces

← All use cases / Health & fitness

ToolClaude

A runner who expected their first 10K to take 70 to 80 minutes finished in 57:22, so they went over the race data with Claude as if it were their coach. It helped them adjust their pace and heart rate, running started to feel easier, and 12 weeks later they ran their second 10K in 54:53, two and a half minutes faster.

Prompt to copy
I recently ran a [DISTANCE] race in [TIME]. Here's my data: [PASTE SPLITS PER KM OR MILE, AVERAGE AND MAX HEART RATE IF YOU HAVE THEM, AND HOW IT FELT]. I usually run [NUMBER] times a week, mostly at about [USUAL PACE]. Based on this, suggest sensible paces for easy runs, steady runs and harder sessions, and explain why. Then outline a simple [NUMBER]-week plan building towards my next race on [DATE], with one rest day after any hard session. Keep it realistic for a recreational runner.

Replace the [PLACEHOLDERS] with your own details before you send it.

How to try it

  1. Export or copy your race data: the time per kilometre or mile, plus heart rate if your watch records it.
  2. Add a few lines on how the race felt and what your usual training week looks like.
  3. Ask Claude to read the data like a coach would: where you started too fast or slow, and what your result suggests your easy and harder paces should be.
  4. Ask for a simple plan towards your next race, and adjust it to fit your real week.
  5. After a few weeks, share new run data and ask whether the paces still look right.

Why this works

Many recreational runners have never tested themselves properly, so they train at paces based on a guess. This runner’s estimate came from a single training run and turned out to be far too cautious. A race gives you a real number to work from, and having it read back to you in plain language turns a time on a screen into training paces you can use. For a gentler starting point, see our guide to training for your first 5K with AI.

Tips to get a better result

  • Include how you felt in each part of the race, not just the numbers; it helps explain a slow finish or a fast one.
  • Ask for pace ranges rather than single numbers, as hills, heat and tired legs all change things.
  • Keep most of your running easy and conversational, and add one harder session a week to begin with.
  • To fit strength work around your running, try our guide to building a workout plan with AI.

Common mistakes

  • Trying to hit faster paces on every run straight after a good race, which is a quick route to injury.
  • Treating heart-rate targets from a chat as medical advice. If you have a heart condition or other health worries, talk to a doctor first.
  • Comparing your plan with elite runners’ schedules rather than building from your own recent data.

Try a variation

Here’s a variation of our own to try next:

Here are my last four weeks of runs: [PASTE DATE, DISTANCE, TIME AND HOW IT FELT]. Am I running my easy runs too fast? Show me which runs were in my easy range and which weren't, and suggest one change for next week.

Related

Read the full guide →

More health & fitness use cases

Prepare for a doctor’s appointment with AI (without self-diagnosing)

A Cedars-Sinai primary care doctor says patients who use AI to organise symptoms and prepare questions often arrive better prepared. He recommends using it to prepare for the visit, not to replace it, and to avoid sharing identifying details.

Full guide →
Source: Cedars-Sinai

Check your weekly gym sets are balanced across muscle groups with Grok and Claude

One lifter typed out a four-day workout split and asked Grok to total the weekly sets for each muscle group, then had Claude turn the numbers into a bar chart. Grok then redesigned the plan for equal volume across muscle groups, and they checked the result by hand, suggesting it could help newer coaches review plans.

Full guide →

Build a half-marathon plan around a training method you like with ChatGPT

After using a standard plan for his first two half marathons, one runner switched to plans ChatGPT built around the 'Norwegian singles' approach. With more consistent weekly mileage, his time improved from 1:46 to 1:37, and he keeps adjusting the weeks as his training goes.

Full guide →

Reference: X @nvaliotti