A new European law will soon require businesses to tell people when they’re talking to an AI instead of a human. It takes effect August 2. If you run a chatbot, an intake bot, or an AI quiz on your site, you may have seen a headline warning that you now have a legal problem. The AI disclosure requirements for coaches are real, and the date is real. For most coaches reading this, though, the law probably doesn’t reach you the way that headline implies.
That gap between the scary version and the honest one is exactly why this deserves ten minutes of your attention. Knowing whether a law touches you is its own kind of relief. And the habit underneath it, telling people plainly when they’re dealing with a machine, is good practice no matter where you or your clients live.
What the new rule actually requires
The law is the European Union’s AI Act, and the part that matters here is Article 50. It covers transparency. Starting August 2, 2026, when an AI system is built to talk directly with a person, that person has to be told they’re dealing with a machine. The official text puts the core duty on the companies that make these systems. Providers must ensure people are, in its words, “informed that they are interacting with an AI system”, unless that’s already obvious from the context.
There’s a second piece. AI-generated content, the images, audio, video, and text a tool produces, has to be marked as artificially generated in a machine-readable way. And a caution before you lean too hard on that word obvious. The European Commission published draft guidance in June 2026 that reads the “obvious from context” exception narrowly. The question it asks is whether a typical person in your audience would clearly recognize they’re dealing with AI without being told. For an intake bot that greets visitors like a friendly assistant, that’s a hard case to make. A visible label solves it.
Whether it actually applies to you
Before you change a thing, find out whether the law reaches you at all. This is where most of the panic falls apart. The AI Act isn’t about where you live. It’s about where your AI’s output lands. The rules reach providers and deployers based outside the EU when, in the text of the law’s scope, the output produced by the AI system is used inside the Union.
So run the honest test yourself. If your clients and site visitors are all outside the EU, and nobody in Europe is chatting with your bot or taking your AI quiz, Article 50 doesn’t bind you right now. If you get real traffic from Europe and those visitors touch a client-facing AI tool, you’re arguably in scope, because the tool is putting AI output in front of someone in the EU.
There’s a third point worth considering. No regulator has gone after a solo coach or a small practice over this. Enforcement attention sits on large platforms and the firms that build the AI, not the one-person business running a chat widget. The penalties the law authorizes are steep, up to 15 million euros or 3% of worldwide turnover, but that’s what the statute permits, not what’s coming for your quiz.
The AI tools in your funnel worth checking:
If you’ve already added AI tools to streamline your coaching business, this is the moment to walk your own funnel and notice where a machine speaks for you. The touchpoints that could trigger a disclosure duty are the client-facing ones.
- Website chatbots and live chat. If an AI answers questions on your site, a first-touch note that it’s an assistant, not you, is the whole ask.
- AI intake and qualification bots. A bot that screens leads before a call is exactly the kind of direct interaction the rule has in mind.
- AI quizzes and assessments. If your quiz talks back conversationally or writes up a personalized result, both the interaction rule and the content-marking rule can apply.
- AI-generated marketing content. Images, video, and copy a tool generates fall under the marking rule for synthetic content.
- AI voice agents. A voice that books or screens callers has to make its AI nature clear in the channel people actually hear.
One caution on that list. No regulator has ruled on coaching tools specifically. This is the general rule read with common sense against the tools coaches really use, not a verdict handed down about our field.
Don’t assume your vendor covers you
It’s tempting to figure your chatbot company takes care of all this for you. Sometimes they do. Often they don’t, at least not completely. The law splits the duty. The provider who builds the system carries part of it. You, the deployer who puts it in front of your clients, carry the rest, including whether that disclosure actually shows up clearly on your page. A badge tucked inside a settings menu isn’t the same as a first message that says you’re chatting with an assistant.
Ask your vendors a plain question. What does the tool disclose, when does it say it, and does switching that on fall to you or to them? Then confirm it’s actually turned on. The answer takes one support email, and it moves the risk off your shoulders and onto a stack you can see.
What to do before August 2:
None of this needs a lawyer for a typical practice. It needs an afternoon. Walk through these four steps.
- Inventory your client-facing AI. List every tool a visitor or client can interact with, from the chat widget to the quiz to the scheduling bot.
- Check who your audience is. Open your analytics and look for EU traffic. If Europe never shows up, this law isn’t your priority today.
- Turn on plain disclosure. Wherever a tool talks to people, make sure it says it’s AI at the first interaction, in words a tired visitor would catch.
- Ask your vendors the coverage question. Find out what each tool discloses and who’s responsible for switching it on, then verify.
Why disclosure is worth doing anyway
Even if the EU never touches your practice, the direction of travel is hard to miss. AI adoption has been one of the defining coaching trends of 2026, and the rules are catching up on every side. California has required bot disclosure in commercial chats since 2019. Colorado’s new AI law takes effect in January 2027. Utah already asks businesses to disclose AI use when a consumer asks. The patchwork is spreading, and every piece of it points the same way.
There’s a deeper reason than compliance, and it’s the one that should actually move you. Coaching runs on trust. A client who later realizes the warm assistant that qualified them was a bot feels a small betrayal, even when nothing was hidden on purpose. Telling people up front costs you nothing and protects the thing your whole practice is built on. If you help others adopt these tools yourself, as an AI transition coach does, modeling clean disclosure is part of the credibility you’re selling.
The tools you use will keep changing faster than any of us can track. The principle underneath them won’t. Be honest with the people you serve about when they’re talking to a machine, and you’ll stay on the right side of every one of these laws without having to memorize a single one…
