Skip to content
AI Act Checker

EU AI Act · Article 50

How to add an AI disclosure to Zendesk

Article 50(1) of the EU AI Act, applicable since 2 August 2026, requires AI systems that interact with people to be designed so that those people know they are dealing with AI, unless this is obvious. If your website runs Zendesk as a chatbot or AI agent, this page shows where to set its name, labels and first message so the notice appears at the start of every conversation.

AI product: Zendesk AI agents (the former AI agents - Essential level became legacy on 11 May 2026) ·

What Zendesk shows by default

Yes. Zendesk automatically shows a "Generated by AI" label on the AI agent's name in the messaging Web Widget, once per group of messages. On mobile SDKs and social channels the name appears as, for example, "Zara · AI".

How to set up the AI notice

  1. In the AI agents workspace, select your AI agent and click Content > Identity. Set the AI agent name, which users see in the widget; Zendesk advises choosing a name that makes clear customers are not talking to a human.
  2. Click Content > Use cases, filter Categories by System replies and open the Welcome reply. Edit the dialogue for each language so the first AI agent message says the customer is chatting with an AI agent and can ask for a person.
  3. Under Settings > CRM integration > Messaging settings, set an Avatar URL for an image that does not look like a staff photo, and choose the Transfer group for escalated conversations.
  4. Keep Zendesk's AI label switched on. On mobile SDKs and social channels it can be disabled per channel with the showAIDisclaimer API parameter, and Zendesk says you are then responsible for compliance.
  5. Publish the AI agent, then open the Web Widget and check that the welcome reply and the "Generated by AI" label on the agent's name both appear.
  6. Still on a legacy AI agent in Admin Center (supported until 10 December 2026)? Name and avatar are under AI > AI agents > AI agents, then your agent > Settings > Identity.

Example first message

Hello, I'm an AI agent. I can help with orders, deliveries and returns, and I'll pass you to our support team whenever you ask.

Let people reach a human

Add an Escalation block with "Forward to an agent" in the dialogue builder wherever customers may need a person, and edit the Failed escalation reply for times when no one is available.

What Zendesk says about the AI Act

Zendesk's AI Trust article says chatbots are cited as limited-risk AI systems under the EU AI Act, subject mainly to transparency requirements it says it already meets, and asks customers to make their own assessment.

Where the data is hosted

With the Data Center Location add-on, AI agents data can be hosted in the EEA; AI agents accounts created before 10 March 2026 support EEA (Belgium) or US hosting only.

Why the first message matters

Article 50(1) obliges providers to design AI systems that interact with people so that those people are informed they are interacting with AI, unless this is obvious, and Article 50(5) requires that information to be given clearly at the latest at the first interaction. The chatbot vendor is usually the provider and a website operator who embeds the bot is usually a deployer, but the operator controls the bot's name, labels and first message, so that is the practical place to make sure the notice appears. Breaches can be fined up to EUR 15 million or 3% of worldwide annual turnover under Article 99(4).

Read the full chatbot disclosure guide

Sources

This page summarises the vendor's own documentation as of the date shown and is general information, not legal advice. Menu names change, so check the linked sources and have your own obligations assessed where needed.