Your client’s weight has been flat for three weeks. Your instinct says cut calories, add a finisher, change the split. Before you touch anything, it is worth asking whether the data agrees, and that is exactly the read a good prompt gives you fast.
This is educational content for coaches, not medical or nutrition advice for individuals. The prompts here keep AI in a supporting role, flagging and summarizing. They are not built to diagnose injuries, prescribe calories or macros, or replace your judgment or a client’s doctor.
Nutrition and training prompts are plain-English requests that ask ChatGPT or Claude to read a client’s food logs, check-ins, and training history, then summarize adherence, flag patterns, or prepare your options before you change the plan. In fitness coaching, they keep the AI in a review role: it surfaces what the data shows, and you make the decision.
This is Part 2 of our prompt library. The first post, ChatGPT and Claude prompts for fitness coaching data, covered the weekly rhythm of check-ins, client replies, and monthly reviews. This one goes deeper into the two areas coaches most often change when they should hold steady: nutrition and programming.
| What these prompts do | Turn a client’s real nutrition and training history into a fast second read before you change a macro, break a plateau, or program around pain |
| Who they are for | Fitness coaches and personal trainers who program nutrition and training themselves |
| What you ask the AI to do | Summarize, compare, and flag patterns, never diagnose, prescribe macros, or set training load |
| What is required | Your check-ins, food logs, and training history in one place, with ChatGPT or Claude connected to that data or pasted in |
| What it changes | You decide from the full picture instead of your gut, without gathering every screenshot by hand |
| Cost | Works with a free ChatGPT or Claude account, and the prompts are free to copy |
Here is what this post covers:
- How Connected Data Makes Nutrition and Training Prompts Work
- Nutrition and Adherence Prompts for Fitness Coaches
- Plateau Diagnosis Prompts
- Fatigue, Deload, and Recovery Prompts
- Programming Around Pain and Physical Limits
- Guardrails for Nutrition and Injury Prompts
- Where Assistant Coach Fits
How Connected Data Makes Nutrition and Training Prompts Work
Every prompt below leans on one thing: the AI reading a client’s real history. A plateau question only works if ChatGPT or Claude can see six weeks of food logs, check-ins, and training at once, not one pasted screenshot.
You can paste that context in each time, which works but gets tiring when every review starts with ten minutes of gathering. Or you connect your coaching platform to ChatGPT or Claude so it looks up the client itself, which our guide to connecting ChatGPT and Claude to your fitness coaching data walks through.
Connected data is what makes these prompts possible. Assistant Coach is the only fitness coaching platform we could verify that lets coaches connect their own ChatGPT or Claude account to the check-ins, food logs, and training history they already keep in the platform, so a plateau or fatigue prompt reads the real record instead of a screenshot you pasted. We checked public docs for Trainerize, TrueCoach, Everfit, FitBudd, MyPTHub, 1FIT, Kahunas, PT Distinction, and Carbon.
Nutrition and Adherence Prompts for Fitness Coaches
Here is the trap with nutrition: when a client stalls, it feels like the plan needs to get cleverer. The research points the other way. Protein works across a wide range, roughly 1.4 to 2.0 g/kg/day for most training clients per the ISSN position stand, with muscle gains flattening beyond about 1.62 g/kg/day. A new macro split is rarely the lever. The lever is usually whether the client is hitting the plan you already set. So the first prompt is not about numbers. It is about adherence.
Adherence before macros
Look at {client name}‘s food logs and check-in notes for the last two weeks. Tell me how many days they logged, how consistent their intake was against their current targets, and where they went off plan most often. Do not suggest new calories or macros. Just show me whether adherence is the real issue before I change anything.
Weekly trend, not daily noise
Take {client name}‘s daily weigh-ins for the last four weeks and give me the weekly averages, not the daily numbers. Tell me the direction of the trend and whether the last two weeks are genuinely flat or just noisy. Do not set a new target.
That last one matters more than it looks: a daily weight chart lies constantly, a case we made in why scale weight misleads coaches.
Plateau Diagnosis Prompts
When a client truly stalls, resist the reflex to cut calories on the spot. A long deficit changes the body: energy expenditure drops more than the weight lost would predict, and hunger climbs. This is metabolic adaptation, and it means a plateau is often the body adapting, not proof the plan is broken. So rule out the cheaper explanations first.
Plateau, ordered by likelihood
{client name}‘s weight has been flat for three weeks. Before I assume we need to cut calories, walk through the data: their logged adherence, their step count or activity notes, their sleep and stress mentions, and how they have been measuring. Lay out the most likely explanations for the stall in order, based only on what the data shows. Do not tell me to change calories.
Diet-break readiness
Read {client name}‘s last six weeks of check-ins and food logs. Summarize how long they have been in a deficit, how their hunger, energy, and mood have trended, and whether they have mentioned burning out on the plan. Do not recommend a diet break or a refeed. Just give me the picture I need to decide whether they need one.
The AI does the gathering and sorting. The decision, cut or hold, stays with you.
Fatigue, Deload, and Recovery Prompts
Fatigue is sneaky because it shows up before performance drops. The clearest early sign is effort creeping up at the same working weight, a load that moved easily two weeks ago. Sports scientists frame training readiness as a balance of fitness, fatigue, and outside stress, and a prompt reads that balance across a roster without opening every log.
Effort trend at the same load
Pull {client name}‘s logged sessions for the last three weeks. For each main lift, show whether the reps in reserve or RPE at the same working weight are getting harder week to week. Flag any lift where effort is climbing but the load has not. Do not tell me to deload. Just show me where fatigue may be building.
Readiness scan
Read {client name}‘s check-in notes on sleep, stress, soreness, and motivation for the last two weeks alongside their session logs. Tell me whether the non-training signals line up with any drop in performance. Do not prescribe a lighter week. Prepare me two questions to confirm before I adjust their volume.
The prompt just makes sure you are not the last to notice a client is running on empty.
Programming Around Pain and Physical Limits
Pain rarely announces itself. It hides in a throwaway check-in line, “shoulder felt a bit off on presses,” or a note in the logger next to a lift they cut short. Across thirty clients, those signals are easy to miss, and this is where AI earns its place.
Pain and limitation scan
Read {client name}‘s check-ins and training-log notes for the last month. List every mention of pain, injury, tightness, or a movement they avoided, with the date and the exercise. Do not decide whether it is minor or serious. Just give me the list so I can follow up.
Substitution options to consider
{client name} said barbell back squats bother their right knee. Based on their current program and logged lifts, list a few lower-body exercises that train similar muscles with less knee load, as options for me to weigh up. Do not tell me which to pick, and flag that they should see a professional if the pain is sharp or lasting.
Read that second prompt carefully. It asks for options to consider, not a decision, and it flags that pain can be a medical signal. That framing is not optional: the moment a prompt starts diagnosing or clearing a client to train through pain, you have handed a medical responsibility to a tool that cannot hold it. We wrote about why that line matters in AI will not replace coaches.
Guardrails for Nutrition and Injury Prompts
The same principles keep any AI prompt safe, tightened for the higher stakes of nutrition and injury.
- Ask the AI to summarize, compare, and flag. Never ask it to set calories, macros, or training load, or to judge how serious an injury is.
- Pain that is sharp, radiating, numb, or lasting is a medical question, not a programming one. Route it to a qualified professional.
- Nutrition targets stay with you. The AI can show adherence and trends. You decide the number.
- Always give a time window. “Last four weeks of food logs” beats “recently” in every prompt.
- When the AI sounds confident, ask which log or check-in it read that from. If it cannot point to one, treat the claim as a guess.
Where Assistant Coach Fits
Assistant Coach is a full coaching platform built for solo and small-team fitness coaches. The core is your daily workflow: structured client check-ins, a workout logger with inline video review, meal plan and workout plan builders, goals, notes and todos, a coach website with lead capture, and clean data export when you want it. AI is woven through that workflow rather than bolted on the side, which is what lets the prompts above read a client’s full history instead of a pasted fragment.
When you link Assistant Coach to ChatGPT or Claude, the AI can use the nutrition and training data the platform already keeps together: food and check-in notes, logged sessions, plans, and goals. The AI integration overview covers what it can see and what stays out of scope. A plateau or fatigue prompt is only as good as the data behind it, and connected data means the AI reads the real record.
Frequently Asked Questions
What AI prompts help fitness coaches review a client’s nutrition?
The most useful ones ask ChatGPT or Claude to summarize how consistently a client followed their plan, flag trends in hunger and energy across recent check-ins, and turn daily weigh-ins into a weekly trend. They show you whether adherence is the issue before you change any number. You still make the call on calories and macros.
Can ChatGPT or Claude tell me why my client’s weight loss stalled?
It can list the likely explanations from the data, such as slipping adherence, lower activity, poor sleep, or measurement noise, and put them in order. It should not tell you to cut calories. A stall is often the body adapting to a long deficit, not proof the plan is wrong, so treat the summary as a starting point for your judgment.
Is it safe to use AI prompts for injury or pain in fitness coaching?
It is safe to have AI surface and organize what a client reported, such as every mention of pain or a movement they avoided this month. It is not safe to let AI diagnose an injury or clear a client to train through it. Pain that is sharp, radiating, numb, or lasting is a medical question for a qualified professional.
Should a personal trainer let AI set macros or training load?
No. Nutrition targets and training load are coaching decisions that depend on the client’s history and how they respond, which the AI does not fully know. Use it to show adherence, flag building fatigue, or prepare options, then set the numbers yourself.
Do these prompts work if my nutrition and training data are in different apps?
They work best when a client’s check-ins, food logs, and training history sit in one place, because a plateau or fatigue question needs the AI to read across all of it at once. If your data is scattered, you can still paste each piece in, but you repeat that gathering for every client.
How is this different from your first set of coaching prompts?
The first library covered the weekly rhythm: check-in reviews, communication, and monthly summaries. This post goes into the two hardest craft calls, nutrition and programming, where the urge to change something is strongest and the data most rewards a second read.
Next Steps
Pick one client who is stalled or grinding right now. Run the plateau or effort-trend prompt, then compare the AI’s read against your instinct. If they agree, act with more confidence. If they disagree, you just caught something before it cost the client a month.
Want to try this on a real coaching workspace? Sign up for Assistant Coach free. Every new account comes with a built-in sample client, so you can test these prompts against realistic nutrition and training data before using them with your own clients.
References
- Jäger R, et al. International Society of Sports Nutrition Position Stand: Protein and exercise. Journal of the ISSN, 2017.
- Morton RW, et al. A systematic review, meta-analysis and meta-regression of the effect of protein supplementation on resistance training-induced gains in muscle mass and strength. PubMed, 2018.
- Trexler ET, Smith-Ryan AE, Norton LE. Metabolic adaptation to weight loss: implications for the athlete. Journal of the ISSN, 2014.
- Greig L, et al. Autoregulation in Resistance Training: Addressing the Inconsistencies. Sports Medicine, 2020.
- Assistant Coach. AI integration overview.
ChatGPT and Claude Prompts for Fitness Coaching Data
Why Scale Weight Misleads Fitness Coaches