The EU AI Act's Article 50 deadline already passed. Most GTM teams do not know if they are compliant.

Article 50 requires that any AI system a person can interact with — chatbots, voice assistants, AI sales agents — clearly discloses that they are talking to AI, at first contact, not buried in a footer. It applies regardless of whether the system counts as high-risk under the Act's other provisions.

It applies globally. A US or UK company running AI-generated outbound, or an AI chatbot that reaches a prospect in the EU, is in scope regardless of where the company is based. And unlike the Act's high-risk compliance timeline, which the EU's Digital Omnibus package pushed back to December 2027, Article 50's disclosure duties got no such delay. They were live from day one. Fines reach EUR 15 million or 3% of global turnover.

There is also a detail most compliance checklists undersell: the obligation splits between whoever built the AI system and whoever deployed it. If you have licensed a third-party AI SDR tool or chatbot and configured it for your own use, you cannot assume the vendor's built-in disclosure covers you. You are responsible for verifying it is actually present, and adequate for how you are using it.

Answering "are we compliant" requires first answering "which of our buyer-facing systems even use AI, and who owns that." For most revenue organisations, that second question does not have a fast answer.

AI has entered GTM stacks piecemeal — a chatbot here, an AI-assisted outbound tool there, a voice agent somewhere else — adopted by whichever team found it useful, with no central inventory and no single owner accountable for what it does when it reaches a real prospect.

That is not a legal gap. It is a visibility gap. Article 50 is simply the first time it comes with a specific date and a specific fine attached, rather than staying comfortably theoretical.

Article 50 includes an exemption where a human takes genuine editorial responsibility for AI-generated content before it ships. Regulatory guidance is explicit that a spell-check or a quick skim does not qualify. It has to be substantive review: fact-checking, real revision, actual sign-off.

A lot of AI-generated outbound at scale is built specifically to avoid that kind of per-message human review — that is the efficiency pitch. Which means a meaningful share of AI-personalised cold outreach currently running probably does not qualify for the exemption it might be assumed to have.

Getting genuinely compliant is not primarily a legal exercise. It is an inventory exercise: which AI-touched systems face a real buyer, who owns each one, and can that ownership actually be demonstrated if asked.

Most revenue leaders we talk to cannot answer that today — not because the requirement is unclear, but because nobody mapped it. That is the same gap that shows up whenever a GTM org tries to scale AI without deciding, in advance, which process is authoritative, which data is trustworthy, and who is accountable when something goes wrong.

If this is showing up in your GTM system, the Lynr team can diagnose the gap and map the highest-impact fix — Signal is delivered in 5 working days from confirmed kickoff, provided the agreed scope, access, evidence sources and stakeholder availability are in place. Start with Signal or book a 20-minute conversation.