Turn a training log into a progress review and next block
Trainers sit on a pile of tracked data — loads, reps, bodyweight, adherence — and rarely have time to synthesize it into a clear progress story. AI is genuinely good at structured synthesis of numbers you give it: what moved, where a plateau is forming, and what to adjust next. The judgment stays yours, and anything that looks like pain or pathology goes to a clinician, not a chatbot.
You are a strength coach analyst. Using only the de-identified training data I provide, summarize progress and recommend adjustments for the next block. The client's goal is {{goal}}; this covers {{timeframe}}. Data (no names or health details): {{training_log}} Produce: 1. Progress vs. goal: what improved, by how much, and what stalled — using only the numbers I gave you. 2. Adherence read: sessions completed vs. planned, and any pattern (e.g., a lift consistently skipped). 3. Next-block recommendations: concrete changes to volume, load, or exercise selection to keep progress moving. Rules: - Use only the numbers I provided. Do not invent lifts, bodyweights, reps, or sessions. If something needed for the analysis is missing, write [MISSING DATA: what you need]. - Do the math carefully and show the before/after for each metric. - If a stall, a drop, or a note suggests pain or a possible injury, do NOT diagnose or explain the cause — write [REFER OUT: recommend medical evaluation] and move on. - Recommendations only. You are advising me, the coach; I decide what the client does.
Fill in your details and the prompt updates live — then copy.
Progress vs. goal (8 weeks): - Back squat: 80kg to 95kg for 5 (+15kg, +19%). Strong, on-goal progress. - Deadlift: 100kg to 110kg for 5 (+10kg, +10%). Steady. - Bench: 60kg to 60kg — flat. Note: 3 bench sessions skipped, which explains the stall. - Bodyweight: 78 to 78.5kg — held, consistent with the goal. Adherence: 21 of 24 sessions (88%), with the misses clustered on bench day. Next block: keep the squat/deadlift progression; move bench earlier in the session to protect it, and drop to a rep target the client can hit consistently. [MISSING DATA: RPE on the top sets would sharpen load calls.]
The full workflow
- Export the client's log from your app and strip names and any health notes
- Run the prompt, then recheck every calculation — AI miscounts volume and percentages
- Turn the recommendations into an actual next block using your own judgment
- Send any [REFER OUT] flag to a physician or physical therapist before continuing to load
Watch out for
Scope: do not diagnose. A regression, a plateau, or a pain note is not yours (or the AI's) to explain medically — refer to a physician or physical therapist for evaluation.
Privacy: de-identify the log. No names, dates of birth, or health conditions in a consumer chatbot, which retains and trains on inputs by default.
Verify the math. Chatbots confidently miscalculate totals, percentages, and progression — the numbers in a client-facing summary have to be right.
Where this comes from
Every use case on this site is grounded in real reports from working personal trainers — not invented by us.