Should You Tell Clients You Use AI? (And What to Say When They Ask)

You've written the program. It's good, and you know it's good, because you spent twenty minutes reworking what the AI gave you: cutting the volume, swapping two exercises around your client's shoulder, adding the deload she's going to need in week three.

Then you hover over send, and something small and uncomfortable happens.

What if she asks?

Most coaches using AI have had this moment. Not guilt exactly. You know the work is yours. But there's an awareness that you've never decided what you'd say, and a question you haven't answered is one you'll answer badly under pressure.

So this article answers it. Not with "be transparent," which sounds complete and tells you nothing, but with a rule for when disclosure actually matters, the specific words for the four situations where it comes up, and the one part of this that most coaches haven't thought about at all.

The Question Behind the Question

When a client asks "did you use AI for this?", they're almost never asking about software.

They're asking one of three things, and which one it is changes your entire answer:

• "Am I getting what I paid for?" They paid for your expertise. They want to know it's in there.
• "Do you actually pay attention to me?" This is the anxious client. Their real fear is being a row in a spreadsheet.
• "Is this safe?" Usually from clients with an injury, a condition, or a history of being handed something generic that hurt them.

Answer the software question and you'll miss all three. A coach who responds to "did you use AI?" with a lecture about drafting tools has technically told the truth and completely failed the client.

Here's the thing: your client isn't running an ethics audit. They're checking whether the relationship is what they think it is. That's a much easier question to answer well, because the answer is yes.

The Rule: The Line Isn't AI, It's Judgment

Every workable disclosure policy I've seen comes down to one distinction, and it isn't between AI and no-AI.

It's between reviewed and unreviewed.

If AI drafted something and you applied real coaching judgment to it, checking the volume against what this client recovers from, tracing their constraints through every week, changing what needed changing, then the program is yours. AI held the pen. You made the decisions. That belongs in the same category as your template library, your spreadsheet, and the certification textbook on your shelf.

If AI drafted something and you passed it along without that pass, the program isn't yours. You sold expertise and delivered a generated document. That's the thing worth being uncomfortable about, and no disclosure statement fixes it. (If you're not sure your review pass is catching enough, the mistakes AI makes in programs are predictable enough to check in about five minutes.)

That gives you a clean rule:

Disclose your process any time a client asks. Disclose proactively when AI touches their personal data, or when it touched a decision you didn't review.

The second half is the one that matters. Almost every coach reading this is fine on the first half. The trouble starts when "I'll review it" quietly becomes "I skimmed it" during a busy week, because now you have something to hide, and that's a workflow problem in an ethics costume.

What You Don't Need to Announce

There's a version of transparency that's really only anxiety, and it makes clients trust you less.

You don't proactively disclose:

• That you used AI to help word a check-in response you thought through yourself
• That you started from a program template (AI-generated or otherwise) and adapted it
• That you used a calculator, a spreadsheet formula, or an app to work out a load progression
• That you drafted your onboarding email with help and edited it

Nobody expects a coach to hand-calculate every number and write every sentence from a blank page. Your client doesn't need to know which parts of your process are manual any more than they need to know which parts of your kitchen involve a food processor. What they're buying is the outcome and the judgment behind it.

Over-disclosing has a real cost. A coach who opens every program with "I used AI to help draft this, but I reviewed it carefully!" has planted a doubt that wasn't there and made themselves sound defensive about their own work. If the review was real, present the program as your work. Because it is.

What You Do Need to Say Out Loud

Four situations flip this.

A client asks. Always answer honestly and specifically. Vagueness here is worse than the disclosure would have been.

Their personal health data is going into a tool. More on this below, because it's the part most coaches have skipped entirely.

You've explicitly marketed the opposite. If your sales page says "every program written personally by me from scratch," you have a positioning problem to fix rather than a disclosure question to answer. Change the page.

Something got through unreviewed. If a client flags an exercise that shouldn't have been in there and you know why it happened, own the process. "That shouldn't have made it through my review. That's on me, and here's the fix" keeps a client. Deflecting loses one.

What to Actually Say

Scripts, because "be transparent" isn't one.

When a client asks casually: "Do you use AI for this stuff?"

"Yeah, for drafting. I use it the way I use a template. It gets the first version on paper fast, and then I do the work of fitting it to you. Your program went through about three rounds of changes before you saw it: the volume was too high for your schedule, and I swapped the overhead work because of your shoulder. That part isn't something AI can do. It doesn't know you."

Naming two specific changes you made does more for trust than any paragraph about your philosophy.

When a client asks suspiciously: "Did AI write this?"

Don't get defensive, and don't over-correct into a speech.

"AI drafted it, I rewrote it. Happy to walk you through what I changed and why. Genuinely, it's a good conversation to have about your programming."

Then actually offer it. A client who takes you up on that comes out more confident than they went in, because they get to watch you think.

When the anxious client asks: "Am I getting a generic program?"

This one isn't about AI at all. Don't let them steer it there.

"No. Here's what's specific to you in this block: [three things]. If any of that stops fitting (your schedule changes, the knee flares up), tell me and we change it. That's the part you're paying for."

When you want to get ahead of it in onboarding

One line in your welcome doc, stated plainly, removes the whole issue before it starts:

"How I work: I use AI tools to speed up drafting and admin so I can spend my time on the parts that need a coach: reviewing your programming, adjusting it as we go, and properly reading your check-ins. Everything you receive has been through my review. If you'd ever like to know how something was put together, ask me."

Coaches who put this in onboarding tend to find the question stops coming up. You've already answered it, and you've framed AI as something you use on their behalf, which is accurate.

What not to say, in any version

• "It's just a tool." Dismissive, and that "just" is doing work you don't want it to do.
• "Everyone uses it." Deflection. They asked about you.
• Anything that shades into "no" when the honest answer is "yes, for drafting." A client who later senses they were misled will not remember your careful phrasing.

The Part Most Coaches Haven't Thought About

Disclosure gets all the attention. Client data gets almost none, and it's the bigger exposure.

When you paste a client's intake form into a chatbot, you're sending someone else's health information (injuries, medications, sometimes medical history) to a third-party company. Your client didn't consent to that. Most likely they've never considered it, which is exactly why it's your call to make carefully rather than theirs to discover later.

You don't have to stop using the tools. You do need a practice:

• De-identify by default. Initials or "the client," not full names. AI doesn't need to know who they are to program for them.
• Summarize instead of pasting. "42F, prior ACL reconstruction, cleared for loading, three sessions a week" carries everything the model needs. Their intake PDF carries a lot it doesn't.
• Keep medical documents out. Physio notes and doctor's letters don't go in a chat window.
• Know your tool's data settings. Consumer chat products and business accounts treat your inputs differently. Ten minutes reading the settings page is worth more than any disclosure statement you'll write. (Choosing tools honestly covers this ground.)

Do these four things and the data question mostly disappears. Skip them and you have a real problem running quietly in the background, one that transparent communication won't touch, because it isn't a communication problem.

The Test That Settles It

If you want one question to run all of this through:

If your client could watch your entire workflow, would they feel misled?

Not whether they'd be surprised. Plenty of clients would be surprised by how any professional works. Misled is the higher bar: would they conclude they'd been sold something different from what they got?

Draft with AI, review it properly, adapt it to the person, stay honest when asked, and they'd feel fine. Some would be impressed you're using your time that well. Send unreviewed output to someone paying for your expertise and they'd feel misled, and they'd be right.

You already know which side of that line you're on. Most coaches who worry about this question are on the right side and haven't given themselves permission to say so out loud.

Frequently Asked Questions

Is it ethical to use AI to write client programs?

Yes, when you review and adapt the output. The ethical question isn't which tool produced the first draft. It's whether the coaching judgment your client paid for is present in what they receive. A program you've checked against the client's recovery capacity, constraints, and history is your program. One you forwarded without reading isn't, regardless of what wrote it.

Do I have to tell clients I use AI?

There's no legal or certifying-body requirement to volunteer it in most places, but you should always answer honestly when asked, and you should disclose proactively in two cases: when their personal health data is going into a third-party tool, and when your marketing has claimed everything is written from scratch. Many coaches find that one plain sentence in their onboarding documents removes the awkwardness entirely.

What if a client reacts badly to finding out?

Usually the reaction is to feeling deceived, not to the tool. Answer the real concern: show them specifically what you changed and why, and offer to walk through their programming. Clients who see the judgment behind the draft typically end up more confident than before they asked. If someone still isn't comfortable after that, it's a fit conversation, and it's better to have it openly.

Is it safe to put client information into ChatGPT or Claude?

Not in raw form. De-identify before you paste: use initials, summarize the relevant details rather than uploading intake forms, and keep medical documents out of chat windows entirely. Check your account's data settings too, since consumer and business tiers handle your inputs differently. A summarized, de-identified client picture gives the model everything it needs to help you program.

Should I mention AI on my sales page?

Only if it's genuinely part of your pitch. Some coaches position "AI-assisted, coach-reviewed" as a feature and it lands well with clients who value responsiveness. What matters more is the reverse: don't claim hand-written-from-scratch programming if that isn't what you do. That's the claim that turns a normal workflow into a credibility problem later.