MY SYSTEM · STEP 5
AI for Golf Amateurs as an Analysis Tool
How Claude and ChatGPT Turn My Golf Data into Training Plans. Six Concrete Use Cases from My Senior Golf Everyday — with the Prompts That Really Work.
With its own data, AI goes from hype to tool
Three years ago, I would have needed a coach for my lap analysis. Today, I get a stroke-by-stroke analysis in 20 minutes — for free, anytime, as often as I want.
The crucial difference compared to „AI asks me for golf tips”: I bring my own data. Garmin R10 CSV, Bebrassie export, Excel spreadsheets from three years. This turns a standard answer into an analysis of my specific weaknesses.
Prerequisite — and this is the uncomfortable truth: Without its own data, AI in golf is just small talk. With its own data, it becomes a tool you have in your hand every day.
What I do with AI specifically
Round analysis
I export my Bebrassie round as a PDF, upload it, and get a shot-by-shot Strokes Gained evaluation. Example 06/12/2026 Wendlohe C: SG -2.23, PW as the biggest score killer with Σ -2.90. Without AI: 2 hours of Excel. With AI: 20 minutes.
Range Data Analysis
4.063 R10 shots in an Excel file. Question for AI: „Which club has the smallest dispersion between 80 and 120 meters?” Answer in 30 seconds: I8 with σ 7.53 m over the last 4 weeks. This is how you find club specialists in your own pocket.
Training Plan Generation
KW25 Plan fully generated from AI: Three training focuses derived from the Course-C round — PW distance control, Lag putts 4-8m, Course management water. 8 PowerPoint slides, ready in 15 minutes. Beforehand: a weekend of work.
Weakness diagnosis
For three rounds in a row, the AI showed me: Pitching wedge is your biggest score killer (confirmed 3 times). I wouldn't have seen this with a coach — the AI objectively compares strokes gained across all clubs, without gut feelings.
Pre-Round Strategy
Before the Wendlohe C+B on June 16: KI received my yardage book and my club distances and created a hole-by-hole cheat sheet with recommendations for each shot. Result: Fairway-in-regulation improved from 25 (%) to 57 (%). That’s how data-driven course management works.
Study research
USGA Distance Insights, TrackMan Reports, BJSM Sports Medicine: The AI reads the PDFs and filters out what is relevant for 75-90 mph swing speeds. This is how the magazine articles are created, without me having to sift through 200-page studies myself.
In five steps to a robust analysis
For „nice output” to become real improvement, discipline in the process is needed. This is how I proceed.
Clean up data
CSV from R10, PDF from Bebrassie, Excel spreadsheet from three years. Before uploading: German decimal separators, meaningful column names, irrelevant rows removed. Garbage in, garbage out — this also applies to AI.
Ask specifically, not openly
„Analyze my round” is garbage. „Calculate Strokes Gained per shot against scratch, show the five weakest shots, derive three training focuses” is an analysis. The prompt determines the quality.
Review the answer critically
AI hallucinates. I double-check calculations with sample checks. For recommendations, I ask, „What studies are you basing this on?”—and I check the sources. Trust is good, but sampling is better.
Translate in training task
„PW is too short” isn't enough. „PW distance control 80m: 30 shots in the 8m circle around 80m” is an assignment. The AI helps with translation — but the assignment needs to become measurable.
Retest after 2–4 weeks
Pre-test, train, post-test. Has the σ-variation decreased? Are the Strokes Gained better? Without re-testing, AI is just entertainment — with re-testing, it becomes an improvement tool.
What you can really copy
I've been using these four prompts for months. They aren't marketing; they provide real analysis.
Stroke-by-stroke analysis with Strokes Gained
Regret comparison by dispersion
Derive weekly plan from weaknesses
Sports Medicine for Senior Golfers Filter
What AI CAN'T do in Golf
- Analyze swing video. Even current AI models cannot reliably detect a path error or attack angle from a smartphone video. For that, you need the R10, TrackMan, or a coach.
- Read the putt lines. Topography, grain, the green's condition – that's work based on experience and gut feeling, not data points.
- Emotion in Competition. Maintaining concentration after a 3-putt, pressure on the 18th hole, ignoring pain — no AI coach.
- Train directly. You have to make the swings yourself. The AI only plans what you do.
- Avoid hallucinations. When the data is unclear, AI makes up numbers. Sampling and source checking are mandatory.
AI doesn't replace a coach — it replaces the Excel weekend that many senior golfers don't have.
Claude · ChatGPT · and Specialists
My main tool. Reads CSV/Excel/PDF with confidence, creates training plans, builds PowerPoints. Strengths: in-depth analysis, honest answers about limitations.
For quick answers and web searches. Often makes up numbers — so check critically. Strengths: broad training knowledge, good language skills.
Garmin Golf App (R10 analysis), Bebrassie (course tracking), TrackMan (on the range). They provide the data, the AI processes it.
German 
English 