How Yext Chat and Search Meet the EU AI Act

TL;DR: The EU AI Act's Article 50(1) transparency deadline hits August 2, 2026, requiring AI systems that interact directly with people to disclose they're not human. Yext's Chat, Search, and certain store locator Pages already ship this disclosure out of the box, so most Yext customers won't need to do anything within the platform to comply.

What's changing on August 2, 2026?

Starting August 2, 2026, the European Union (EU) AI Act's Article 50(1) requires any artificial intelligence (AI) system that talks directly to a person (a chatbot, an AI-generated search result) to tell that person they're talking to AI, not a human. This applies to any brand whose AI experience reaches someone in the EU, no matter where the brand is headquartered.

What do we need to do as a Yext customer?

The short answer: the disclosure is already built in. Yext built this before the regulation forced the issue.

If you run Yext Chat, Yext Search direct answers, or a Yext-powered store locator, you're likely already covered. The disclosure ships out of the box. You don't need to design or build anything new.

Two things to confirm before August 2, 2026:

• Check your SDK version. The signposting ships in the current Search and Chat SDKs and UI components. If you embed either one on your own site and it reaches people in the EU, make sure you're on the current release.
• Decide if the default fits. The out-of-the-box signposting works as-is. If your brand or legal team wants different wording or a different look, you can customize it.

The rest of this post covers what the regulation requires, what disclosure needs to look like, and how the built-in signposting works.

What the EU AI Act's Article 50(1) deadline actually requires

The EU AI Act (Regulation (EU) 2024/1689) is the world's first broad law governing artificial intelligence. It took effect on August 1, 2024, but its obligations are phasing in on a schedule.

Prohibited practices and AI-literacy requirements landed in February 2025. Rules for general-purpose AI models followed in August 2025. The next major milestone lands on August 2, 2026.

That milestone includes Article 50 – the transparency and labeling requirements. Article 50(1) covers a narrower, concrete obligation. When an AI system interacts directly with a real person, that person has to be told they're interacting with AI, not a human.

Say your chatbot answers a customer's question, or your AI search returns a direct answer instead of a list of links. Either way, the person on the other end needs a clear signal that AI is doing the talking.

Who this applies to, including brands outside the EU

The EU AI Act has a broad reach. A U.S.-headquartered retailer with EU storefronts, or a global brand whose AI chat widget is available to EU visitors, is in scope. Location doesn't exempt you. The deciding factor is whether someone in the EU can interact with the system.

What "AI disclosure" means in practice for a chatbot or AI-driven search

Conversational AI agents, like Yext Chat, are exactly the kind of system Article 50(1) has in mind. In practice, this comes down to a clear, visible signal at the point of interaction. An icon, a label, or a short note that tells the person they're talking to AI. It has to register before or as the conversation starts, not buried in a terms-of-service page three clicks away.

For a chatbot, that might be an icon next to the first message.

For an AI-generated search result or a direct answer, it might be a label on the response itself.

Why highly regulated brands have the most to lose from getting this wrong

For brands in financial services and healthcare, this isn't a new category of risk so much as a new item on an already-long compliance list. These industries already answer to regulators who take disclosure seriously, and an AI-transparency miss compounds exposure that's already higher than the average.

A generic, bolted-on disclaimer is its own kind of red flag when it doesn't match how the rest of the brand's compliance program is run. Inconsistency reads as an afterthought, not control.

How Yext's Chat, Search, and Store Locator signposting works

The signposting is designed to be easy to deploy and easy to make your own. It's configurable across three dimensions:

• The signposting icon itself
• The hover state a person sees before they interact
• The title and description shown when someone clicks or taps the signposting element

That's an AI data governance decision Yext already made for you: the signposting ships as a configurable layer on the product itself, not a separate compliance add-on.

That's really the shape of this one: a regulatory deadline that would otherwise mean new engineering work has already been solved at the product level. The part that's left is confirming your SDK version and deciding whether the default experience fits your brand as-is.

This is what it looks like when governance is built into the product up front, rather than bolted on after a deadline forces the issue.