Skip to content

Article

Should you tell clients you use AI to write their programs? Yes, and here is how

Most advice on this question is either a legal hedge or a shrug. This is a coach's position with the exact words to send, the objections you will hear, and a one-page policy for a facility.

By RaiNGE product teamUpdated 10 min read

Written for coaches and gym owners deciding how to talk to clients about AI-drafted programs.

Short answer

Yes. Tell clients in one plain sentence at onboarding and again in the program notes: AI drafts, and you review, edit, and approve every program. Trust is the product in coaching, your review is the value, and a client who finds out later reads it as concealment. This is RaiNGE's view, not legal advice.

  • Say it twice: one line at onboarding, one line in the program note.
  • Name your review in the sentence. The words that matter are "I approve every program."
  • Answer the data question with what you do, and never promise what you cannot verify.
  • Rules differ by place. A facility owner should check with their insurer or counsel.

Before you tell clients about a tool, run it on your messiest client and score what it does with the injury note and the missed week. Test an AI tool on your hardest client

On this pageThe position: tell them, and say what your review does

The position: tell them, and say what your review does

This is RaiNGE's opinion, argued the way a coach would argue it at a staff meeting. It is not legal advice, and it says nothing about what any law requires. Disclosure rules differ by place, and the note at the end of this page says who to ask. The argument for telling clients rests on three reasons.

  1. 1

    Trust is the product

    Clients do not pay you for exercise names. They pay a person who knows their knee, their schedule, and their bad weeks. Anything that changes how their program gets made is part of what they are trusting you with, so say it.

  2. 2

    Your review is the value

    An AI draft is only a draft. What turns it into a program is the coach who reads it against the client's notes, changes what is wrong, and approves it. If you hide the draft, you hide the work you did on it.

  3. 3

    Finding out later reads as concealment

    Say a client learns from a friend, a screenshot, or a leaked tool name that their program was AI-drafted. The program may be excellent. Their next question is still what else you did not mention, and you are now answering that on the back foot.

Say it in two places. The onboarding line sets the expectation before the first program. The program note repeats it when the client is looking at the work, which is the moment the question actually occurs to them. One mention at onboarding is forgotten within a month.

The disclosure ladder: five levels, what each costs and what it buys

Disclosure is not all or nothing. These five levels run from silence to full transparency. Pick one on purpose, write it down, and apply it to every client.

Five levels of disclosure for AI-drafted programs, from saying nothing to showing the draft and the edits.
LevelWhat you doWhat it costs youWhat it buys you
0. Say nothingUse AI for drafts and mention it to no oneNothing today. Exposure to the finding-out problem on any given dayNothing. Every question the client asks later, you answer from behind
1. Answer if askedBe honest when a client asks, volunteer nothingNothing until someone asks, then an awkward momentHonesty, but only with the clients curious enough to ask
2. One line at onboardingOne sentence in the intake form or welcome messageAbout 10 seconds of reading per clientThe expectation is set early, and no one can say they were not told
3. Onboarding line plus program noteThe same sentence at the start, plus a line with each new block saying what you changedAbout 2 minutes per block to add the noteThe client sees your review at the moment they are judging the work
4. Show the draft and the editsShare what the AI drafted next to what you approved, with your reasons5-10 minutes per client, and the nerve to be seen editingThe strongest proof that your review is real. Best for skeptical clients and complex injury cases
Five levels of disclosure for AI-drafted programs, from saying nothing to showing the draft and the edits.

Level 3 is the default RaiNGE would recommend: cheap, repeatable, and it keeps your review in front of the client. Use level 4 for clients who ask, for anyone with injury notes who wants to see the reasoning, and for a facility that wants a stronger trust story than a sentence. The guide to explainable AI workout recommendations lists what a good explanation of an edit should show.

Three messages to copy and adjust

Square brackets mark the parts you fill in. Keep your own voice, and cut anything you would not say out loud. Every message names the review, because that is the part clients care about.

One line for onboarding

“I use AI to draft the first version of your program from your goals, history, and equipment, and then I review, edit, and approve every program myself before you see it.”

A short paragraph for the program note

“Hi [first name], your new block is ready. A quick note on how it was built: AI helped me draft the first version from what you told me about your goals, schedule, and [injury or equipment note]. I then went through it line by line and changed [one real change, for example: swapped the walking lunges for step-ups because of your knee]. If you want to know why any exercise is there, ask me and I will tell you. [Coach name]”

A reply to a client who objects

“Thanks for telling me, [first name]. That is a fair thing to raise. Here is exactly how it works: AI produces a first draft, and I decide what stays, what changes, and what you do. You are paying for my judgment on your training, and that has not changed. If you would rather I write your program without AI, I can do that. [Say honestly what it changes, for example: it takes me a little longer, so your next block will land two days later.] Which would you prefer?”

Only promise what is true for you. If you do not change an exercise in a block, say what you checked instead, because an invented edit is worse than none.

Objections and replies, starting with the strongest

The strongest objection is that telling clients lowers the perceived value of your work. Some clients will think less of a program they know a machine drafted. The reply is that the value was never in typing the week. It sits in what you notice, change, and decide, and the disclosure line is where you point at it. A coach who says "I review, edit, and approve every program" is claiming the work that matters.

Other objections are more practical. Here are the seven you are most likely to hear, what the client is really asking, and a reply you can adapt.

Seven client objections to AI-drafted programs, what each one means, and a reply to adapt.
What the client saysWhat they are really askingA reply you can adapt
Is that cheating?Am I getting less than I paid for?No. Drafting is prep work. I read every line against your notes, change what is wrong, and approve it. The decisions are mine, and so is the result.
Will AI replace you?Is my coach going away, or am I paying for nothing?It takes over the prep, so I have more time for your sessions and check-ins. It cannot watch you move or hear that you are having a rough week. That is the job.
Is my data safe?Where does my information go?Say what you do and nothing more: for example, I leave your name and contact details out of any general-purpose AI tool and share only training details. I cannot vouch for how every tool stores information, but I can show you which one I use and its terms.
Can I opt out?Do I get a say?Yes. Tell me and I will write your program without AI. I will be straight with you about anything that changes, such as timing or price.
Who is responsible for my program?If it goes wrong, who answers?I am. I approve everything you train. If anything hurts, tell me, and I decide what changes. The tool does not make that call.
Does it know about my injury?Is this safe for me specifically?It only knows what I give it. I put your injury notes in, and I read every exercise against them before I approve the week. If your notes change, tell me and I will update them.
Honestly, I do not mindNothing. Do not make it a conversationGreat. One line, then move on. The point of the line is that they were told, not that they had a discussion.
Seven client objections to AI-drafted programs, what each one means, and a reply to adapt.

Worked example: rolling it out to 20 clients over two weeks

This is an invented example, not data from RaiNGE or from any coach. Say Dana is a fictional coach with 20 clients on four-week blocks, so about five new blocks land each week. She has not told anyone she drafts with AI. Here is her plan.

  1. 1

    Day 1: write the two lines

    Dana writes her onboarding sentence and her program-note paragraph, using the templates above. She puts the sentence into her intake form and welcome email so new clients see it from now on.

  2. 2

    Days 2-3: tell the existing 20

    She sends one message to all 20 existing clients, in her own voice: “Quick note on how I build your programs. I use AI to draft the first version from your goals, history, and equipment. Then I review, edit, and approve every program myself before you see it. If you have questions or would rather I write yours without AI, just reply and tell me.”

  3. 3

    Week 1: add the note to each new block

    As the five blocks go out, each one carries the program-note paragraph with one real change named. She keeps a one-line log of what she changed and why, which doubles as her level 4 evidence if anyone asks.

  4. 4

    Week 2: answer, count, and adjust

    She replies to every question the same day and logs each one. At the end of the week she reads the log and rewrites any reply that did not land.

Dana plans for about 4 replies out of 20, one in five, mostly "which tool is it?" and "is my data safe?", plus at most one request to opt out. That is a planning number for her calendar. It is a guess, not a prediction from data, and if nobody replies, that is fine too. At about 5 minutes per reply, she reserves 20 minutes across the two weeks. The example claims no outcome for retention, trust, or revenue.

A policy checklist for gym owners

A single coach decides this once. A facility has to decide it for every coach, including the one who starts next month. The AI for gym owners guide covers a full pilot and the wider approval policy. This checklist covers the disclosure part.

Write these down before the first message goes out

  • Who discloses: each coach, in their own words, using the approved sentence. The front desk does not improvise it.
  • Where it is written: the intake form, the welcome message, and the program-note template. One sentence in each, worded the same way.
  • What happens when a client opts out: who writes that client's program, how long it takes, whether anything changes on price or timing, and where the opt-out is recorded on the client's file.
  • What staff may not paste into a general-purpose AI tool: names, contact details, birth dates, payment information, and health notes tied to an identifiable person.
  • De-identification: use a code for the client's name, describe injuries in training terms such as low-back pain 3/10 on hinges, and leave out dates, employers, and locations. The guide to chat-tool workout programs shows a worked prompt.
  • Who answers the hard questions: one named person handles data questions and complaints, so coaches do not invent answers.
  • Review date: put the policy on the calendar 90 days out, and update it whenever you change tools.

RaiNGE is built on the same rule the message relies on: AI output is a draft, and a coach reviews, edits, and approves a program before it reaches a client. The AI workout generator for coaches page shows that workflow.

Common questions

Do I have to mention AI if I only use it for exercise ideas?

Drafting a whole program and asking for three single-leg exercise ideas are different uses. A practical line: if AI-written text or structure reaches the client's program or messages and you did not rewrite it, mention it. If you used it the way you would use a search engine, one sentence in your policy is enough. Check local rules too.

Should I name the AI tool I use?

Say "AI" in the onboarding line, and name the tool when a client asks. Have its data terms to hand when you do. Naming a tool in every message adds little and goes stale when you switch products. Keep one place, such as your policy or welcome pack, where the current tool is listed.

What if a client finds out before I tell them?

Tell them directly, the same day, and own the gap: "I should have said this at the start." Then use the objection reply above and offer level 4, showing the draft next to your edits for their program. Do not argue that they should not mind. The goal is to be the source of the information from now on.

Should my website and marketing mention it too?

Keep the website consistent with what you tell clients. If you disclose in onboarding, do not hide it on your site. Describe your review, such as "I check and approve every program", and avoid selling AI as the benefit or promising results the software cannot deliver. A short line in your FAQ is usually enough.

Next step

Before you tell clients about a tool, run it on your messiest client and score what it does with the injury note and the missed week.