Is your customer chat ready for August 2, 2026?

On 2 August 2026, the EU AI Act — Regulation (EU) 2024/1689 — reaches full application. If there's AI anywhere in your customer chat, the rules concern you too. Here's what the law actually says, in plain language, and why how you use AI matters more than whether you use it.

Last reviewed 10 June 2026

This page is general information, not legal advice. For your own situation, talk to a lawyer.

What the AI Act is

The EU AI Act — Regulation (EU) 2024/1689 — is the EU's common rulebook for artificial intelligence. It's a regulation, not a directive: it applies directly in every member state, including Finland, without waiting for separate national laws.

Its rules phase in over time, and the act reaches full application on 2 August 2026. The basic idea is simple: the bigger the risk an AI use poses to people, the stricter the requirements.

Customers must know when they're talking to AI

Article 50 of the act is about transparency. Its core: people must be told when they're interacting with an AI system, such as a chatbot — unless that's already obvious from the situation.

In practice: if the chat on your website answers customers by itself, your customers need to know that. For most small businesses, this is the part of the act that actually touches daily life — not the heavy high-risk requirements written for things like hiring tools.

Human review changes the picture

The act doesn't treat every use of AI the same way. Transparency obligations differ when content gets substantive human review and a person holds editorial responsibility for it, compared with a machine answering on its own.

The difference is common sense. A chatbot answering by itself is an AI system the customer is conversing with. A reply that a person has reviewed, edited and approved is your business's own message, with AI as the helper — closer to a drafting assistant than a chatbot. What this means for your own disclosures depends on your setup, which is why the note at the bottom of this page matters.

The four risk tiers, briefly

The act sorts AI uses into four tiers:

  • Prohibited: certain uses, such as social scoring, are banned outright.
  • High-risk: AI in areas like hiring or credit decisions — allowed, but tightly regulated.
  • Limited-risk transparency: chatbots and the like — allowed, but people must be told about the AI. Most customer-service chat sits in this tier.
  • Minimal risk: most everyday AI, such as spam filters — no new obligations.

Where työ stands

In työ, your customers always talk to a person. AI only drafts a reply, grounded in your business's own knowledge and tone; you or your team review, edit and send every message. Nothing goes out without a human sign-off, and sensitive replies can require two.

We won't tell you that a tool makes your business compliant with the AI Act — don't believe anyone who does. What we can say is this: työ is designed for the human-oversight approach the AI Act rewards. A person decides, the machine assists, and a record is kept.

Frequently asked questions

Does the AI Act apply to my café?

It can. The act applies to businesses that use AI, not only to companies that build it. If there's AI anywhere in your customer chat, it's worth knowing what role it plays: does the machine answer on its own, or does it help a person answer? Your obligations depend on the use — for a firm answer about your situation, talk to a lawyer.

Do I have to label my chat as AI?

Article 50 requires that people are told when they interact with an AI system such as a chatbot. In työ, your customers talk to a person: AI drafts, and a human reviews and sends every message. What you need to tell customers in your own setup depends on the details — check your situation with a lawyer.

What happens after August 2, 2026?

The regulation reaches full application, and national authorities supervise it. The date is no reason to panic — but it is a good reason to find out now how your tools use AI and who answers for the messages your customers receive.

Does using työ make me compliant?

No — and no tool can. Compliance depends on your whole business, not one piece of software. työ gives you a structure where a person approves every message and approvals leave a record, but checking your own situation is up to you and, when needed, your lawyer.

Sources

Still wondering about something? We'd love to hear from you.

Talk to us about governed messaging