How to Write ChatGPT Prompts for Fitness Coaching

Most guides to using AI for coaching hand you a list of prompts to copy. That is useful, but it runs out the moment you have a question the list does not cover. The more useful skill is writing your own, and it is far simpler than “prompt engineering” makes it sound.

A good coaching prompt is a plain-English request that tells ChatGPT or Claude three things: what client data to read, what to do with it, and what not to do. That is the whole skill. There is no special syntax, no code, and no technical background required. If you can give a clear instruction to a new assistant, you can write a prompt.

This is the last part of our prompt library. The others give you ready-made prompts for check-ins and reviews, nutrition and programming, your coaching business, onboarding, and managing a roster. This one teaches you to write your own.

What this is	A simple pattern for writing your own ChatGPT or Claude coaching prompts
Who it is for	Fitness coaches and personal trainers who want more than a fixed list of prompts
The pattern	Name the data, name the task (summarize, flag, compare, draft), name the limit (do not diagnose or prescribe)
What you need	No technical skill; plain English is enough
What it changes	You can answer any coaching question you have, not just the ones someone pre-wrote
Cost	Works with a free ChatGPT or Claude account

Here is what this post covers:

1. What Makes a Good Coaching Prompt
2. The Summarize, Flag, Compare Pattern
3. Always Name the Client and the Time Window
4. Keep the AI in an Assistant Role
5. Adapt Any Prompt in This Series
6. Guardrails for Writing Your Own Prompts
7. Where Assistant Coach Fits

What Makes a Good Coaching Prompt

The difference between a prompt that helps and one that wastes your time is specificity. “How is my client doing?” gives you a vague, generic answer. A good prompt gives the AI three things to work with:

• The data: which client, and which part of their history. “Sarah’s last four check-ins,” not “Sarah.”
• The task: what to actually do. “Summarize what is trending up and down.”
• The limit: what to leave to you. “Do not suggest changes.”

Put together, that is a prompt: “Read Sarah’s last four check-ins, tell me what is trending up and down, and do not suggest changes.” Specific, useful, and easy to check against the real data.

The Summarize, Flag, Compare Pattern

Almost every useful coaching prompt asks the AI to do one of a small set of jobs. Learn these verbs and you can build any prompt you need:

• Summarize: condense something long into something readable. “Summarize this month’s check-ins.”
• Flag: find and surface specific things. “Flag any week her adherence dropped.”
• Compare: put two things side by side. “Compare his logged weights this block to last block.”
• Draft: write a first version for you to edit. “Draft a check-in reply I can adjust.”
• Prepare: get you ready for a decision. “Prepare the questions I should ask before I change her plan.”

Notice what is missing from that list: diagnose, prescribe, decide. Those are your job. The verbs above keep the AI doing the reading and drafting while you do the coaching.

Always Name the Client and the Time Window

Two small additions make almost any prompt sharper. Name the client, and name the window of time.

“Recently” is the enemy of a good answer. The AI does not know if you mean this week or this quarter, so it guesses. “The last four weeks” or “since we started the current block” removes the guessing and makes the answer something you can actually verify against the data.

The same goes for the client. A prompt tied to one named client (“David’s last three sessions”) gets a grounded answer. A prompt about “my clients” in general gets a generic one, unless you genuinely want the whole roster, which is its own kind of prompt we covered in managing your roster.

Keep the AI in an Assistant Role

The single most important habit is adding a limit. AI will confidently do more than you asked, including handing out medical or nutrition advice it has no business giving. A short “do not” line keeps it in its lane:

• “Do not set calories or macros.”
• “Do not tell me to change the training load.”
• “Do not diagnose anything or decide if it is serious.”
• “Do not send anything; just draft it for me.”

This is not a technicality. It is the line between AI as an assistant that surfaces information and AI as something making coaching decisions it is not qualified to make. We wrote about why that line matters in AI will not replace coaches. Keeping the final call with you is the whole point.

Adapt Any Prompt in This Series

You do not have to start from a blank page. The fastest way to learn is to take a prompt that already works and make it yours. Every prompt in this library follows the same pattern, so adapting one is simple:

1. Find a prompt close to your question, for example a check-in summary from the first prompt guide.
2. Change the client name to yours.
3. Change the time window to the one you care about.
4. Change the focus to your question, keeping the “do not” line.

Run a handful this way and you will stop thinking of it as adapting a template. You will just be asking clear questions, which is all a prompt ever was.

Guardrails for Writing Your Own Prompts

A few habits keep your own prompts safe and useful.

• Always include a “do not” line for anything touching health, nutrition targets, or training load.
• Ask the AI to point to the data behind a claim. If it cannot, treat the claim as a guess.
• Be specific with the client and the time window every time.
• Read every drafted message before you send it. The AI does not know your relationship with the client.
• When an answer looks off, it usually means the prompt was vague, not that the AI failed. Add detail and try again.

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. The prompts you write are only as good as the data behind them, and that data lives in your coaching platform.

When you link Assistant Coach to ChatGPT or Claude, your prompts can read real client history instead of context you paste in by hand. That is what lets you write a prompt as a normal coaching question. The AI integration overview covers what the AI can see and what stays out of scope.

The better your coaching data is organized, the better your prompts work, because a clear question needs clear data to answer it.

Frequently Asked Questions

Do I need to learn prompt engineering to use ChatGPT for coaching?

No. You do not need any technical skill or special syntax. A good coaching prompt is a plain-English request that says what data to look at, what to do with it, and what not to do. If you can write a clear instruction to an assistant, you can write a prompt. The pattern in this post is all most fitness coaches ever need.

What does a good ChatGPT prompt for fitness coaching look like?

It has three parts: the data to read (a client’s last four check-ins), the task (summarize, flag, or compare), and the limits (do not diagnose or prescribe). For example: “Read Sarah’s last four check-ins, tell me what is trending up and down, and do not suggest changes.” Naming the client and the time window makes the answer far more useful.

Why should I tell the AI what not to do?

Because AI will confidently do more than you asked if you let it, including giving medical or nutrition advice it should not. Adding a limit like “do not prescribe” or “do not set calories” keeps it in a supporting role: summarizing and flagging, not deciding. The “do not” line is what keeps you the coach and the AI the assistant.

Can I just reuse prompts other coaches wrote?

Yes, and it is a great place to start. Take any prompt from our prompt library, change the client name, the time window, and the focus area, and it is yours. Once you have run a few, you will start writing your own without thinking about it, because the pattern is always the same.

What is the difference between a vague prompt and a good one?

A vague prompt like “how is my client doing” gives you a vague answer. A good one names the client, the time window, the data, and the task: “Summarize David’s last three check-ins and flag any week his tone changed.” The more specific the request, the more useful and checkable the answer.

Next Steps

Pick one question you have about a real client right now. Write it using the pattern: name the data, name the task, add a “do not” line. Run it, check the answer against what you know, and adjust the wording. That is the whole loop, and after a few rounds it becomes second nature.

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 practice writing prompts against realistic data before using them with your own clients.

References

• OpenAI. Apps in ChatGPT.
• Assistant Coach. AI integration overview.