Do you have to tell clients their programme was written with AI?

Every coach using AI for programming eventually has the same conversation with themselves, usually at 11pm, usually with a client's name in the subject line. Do I say anything? The honest answer is that almost nobody is asking whether it is technically allowed. They are asking whether saying it out loud costs them the client. So let's take the two questions separately, because they have different answers.

What coaches are actually worried about

Read enough trainer forums and the fear sorts itself into three piles. The first is embarrassment: that the client will decide the coach did not really do the work. The second is trust: that the plan will be treated as worth less the moment a machine is mentioned. The third is professional: that admitting it invites a question about what the client is paying for.

Those are fair worries, and two of them have evidence behind them. When FitBudd surveyed coaches in 2026, safety and output quality came top of the concern list, not job loss. What worries coaches is what comes out of the tool, which is a different problem from whether to mention the tool.

The third worry is the one worth taking seriously, because it is the only one that is completely within your control. If the pitch was "I write every plan from scratch, personally, using twenty years of experience", then a client finding out about the drafting is not a disclosure problem. It is a positioning problem that started on the sales call.

Why "I use AI" is the wrong sentence

Here is the framing that causes most of the damage, and it is not the client's fault. "I use AI" describes your tools. Clients do not buy tools, and they do not care about yours any more than they care which spreadsheet your accountant uses.

What they bought was judgement. Someone who knows which session to cut when their knee flares up, who notices the pattern in their check-ins, who changes the plan when their shift pattern changes. That is the thing being paid for, and the drafting is packaging around it.

So the disclosure question gets easier the moment you stop treating it as a confession and start treating it as a description of how you work. Skip "I use AI to write your programme", which describes your tools. The sentence that lands is closer to "I draft your programme in a system that does the arithmetic and the formatting, and I go through every line before it comes to you." Both are true. Only the second one is about you.

That second sentence should also be literally true, not just convenient. If you are handing over output you have not read, you have a bigger problem than disclosure, and no script fixes it.

Where disclosure stops being optional

There are situations where the question stops being strategic, and it is worth knowing them before they arrive.

The first is anything outside your scope. If a client's programme is shaped by a health condition or a rehabilitation protocol, then whoever is making the judgement needs to be a person with the training to make it. That has nothing to do with AI; it is the same reason you refer out. Our own position on this is unusually blunt: injuries, conditions, pre and postnatal screening and anything beyond a trainer's scope stay with the coach, and we build the software around that limit rather than pretending it does not exist.

The second is anything contractual. If you work inside a gym, a franchise or a corporate wellness contract, read the agreement. Some operators now mandate or ban AI tooling outright, and a client-side promise you cannot keep is worse than an awkward conversation at the start.

The third is anything you have explicitly promised. If your marketing says every plan is written by hand, then either stop saying that or stop using the tools. You cannot have both, and the moment a client works it out is the moment you lose something more valuable than the argument.

What to actually say

The version that works tends to be short, specific and slightly boring, which is exactly why it lands. Something like: "Your programme comes from a system I use for the drafting. It does the maths and the formatting, and it flags anything I should look at twice. I go through every line before it reaches you, and I change whatever needs changing."

Then stop talking. The instinct is to fill the silence with reassurance, and that is what makes it sound like a confession. A sentence delivered plainly and left alone reads as a description of a process. The same sentence padded with apologies reads as a problem.

Notice what is not in there. No naming of the system, no discussing how it works, no comparison to other coaches. Clients do not want a product update. They want to know who is responsible for the thing they are following.

The part that is genuinely yours to decide

Two honest positions exist, and both are defensible. Position one is that your process is your business, in the same way a physio does not narrate which software they use for notes. If the plan is reviewed, owned and delivered by you, the tools behind it belong to you. Plenty of excellent coaches work this way, and the clients are no worse off for not knowing.

Position two is that you say it up front, before it can be discovered, because you would rather control the framing than react to it. This is more comfortable for coaches whose marketing leans on the craft, and it survives the difficult client who asks the follow-up question.

What does not work is a third position: saying nothing while implying everything, hoping the subject never comes up. That is the one that turns a routine question into a crisis, and it is the reason the whole topic feels awkward in the first place.

If you want to see what the reviewed output actually looks like before you decide how you would describe it, the page on AI programmes for coaches walks through the pipeline, and you can generate one free plan and read every line of it yourself. That is probably the fastest way to settle the disclosure question, because you cannot describe a process you have not seen.