ChatGPT
AI is here, and now it’s a big part of sport too.
Used well, it’s an incredible resource for athletes and coaches. The real question is: how do we use it properly?
The best use is conversation, not a single answer
For me, the sweet spot is treating ChatGPT like a conversation partner, not a search engine. Athletes get the most value when they ask a question, read the response, then keep digging: what else could it be, what would I test, what data matters, what’s the risk, what’s the next step?
Case study 1:
GI distress in a 70.3
One athlete was dealing with gastrointestinal distress in 70.3 racing. They started by asking why it might be happening, then continued the conversation. ChatGPT didn’t “solve” it, but it did lay out some plausible mechanisms and factors to consider.
The athlete became educated, and then came the best part: they shared the screenshots with me. From there we discussed the likely scenarios and agreed on a few approaches to test.
Why did this work well?
The athlete could explore the issue in their own time and space. No rush - no judgement
The responses were written down, which gave space to absorb and reflect and even go back to later
When they came to me, we started with better questions and better context.
Case study 2:
Quad pain on the run
Another athlete had quad pain during the Ironman run, one of the most common problems in long-course triathlon.
They asked ChatGPT: “What could this quad pain be, and how do I resolve it?”
ChatGPT produced a helpful list of potential causes and practical steps to investigate. Again, the athlete shared the conversation with me, and we are using it as a starting point for a valuable coaching discussion.
This is where AI can strengthen the coach–athlete relationship. It doesn’t replace coaching. It can improve the quality of the conversation.
So what do we learn from this?
I asked ChatGPT directly:
“How does ChatGPT make assumptions on technical issues in triathlon, such as sore quads on an Ironman run?”
The answer was reassuring as a coach and as a human.
(ChatGPT answers below in italics)
ChatGPT:
Does not truly know
Does not diagnose
Does not observe
Infers based on statistical similarity
It is exceptionally good at:
Hypothesis generation
Checklist logic
Risk factor clustering
It is weaker at:
Individual biomechanics
Hidden stress loads
Athlete psychology
Pattern-of-life behaviours
That is exactly where elite coaching lives.
Bottom line for coaches
Think of ChatGPT as a fast, broad, probability-driven assistant that generates starting hypotheses, not final truths.
Then you:
Validate with data
Validate with context
Validate with lived coaching pattern recognition
That partnership is where it becomes powerful.
The key difference between you and ChatGPT
ChatGPT:
Works on population probability
Sees averages
Is optimised for common failure patterns
Cannot feel tension, observe posture, or hear doubt in an athlete’s voice
You:
Work on individual causality
See exceptions
Read micro-signals: how they talk post-race, compliance patterns, stress behaviour, nutrition honesty, training avoidance
Interpret context, not just symptoms
You operate in high-resolution reality. ChatGPT operates in low-resolution probability space.
Conclusion :
The smart way to use it is give it as much detail as possible
Instead of:
“Why were my quads sore on the Ironman run?”
Use:
“Male 35–39. IF 0.74. VI 1.09. 85g/hr carbs. 5:15 bike split. HR drift 7%. Cadence 91. Undulating course. Quads failed at 18km. What are the likely limiting mechanisms, and what would you test first?”
Give it a screenshot of your Training Peaks file !
That’s when ChatGPT shifts from generic answers to more constrained modelling.