AI Comply ShieldAdd an AI notice to Intercom
Short answer: Turn on "Show AI Agent label in Messenger" in Fin's identity settings, and make Fin's intro message say it is AI. Intercom's help centre says the label is off by default for workspaces that deployed Fin after mid-2025, and that customers are responsible for making sure end users know they are talking to AI. Article 50(1) of the EU AI Act has applied since 2 August 2026.
If you use Intercom, the AI in your Messenger is Fin. Intercom has built disclosure settings for Fin — but for many newer setups they are not switched on. This page shows how to check yours, and where to add a clear notice for EU visitors.
When Article 50(1) can apply to your Intercom chat
Article 50(1) of the EU AI Act has applied since 2 August 2026: people interacting with an AI system should be told so, unless it is obvious. It applies based on where the visitor is, so any Intercom workspace with EU visitors talking to Fin is affected.
Fin is unambiguously an AI system. It writes its own answers, and Intercom lets you rename it and give it your own avatar — which is exactly what can make it look like a human colleague to a visitor.
Intercom's own help documentation is direct about this. Fin used to show an "AI Agent" label automatically; Intercom now lets you turn that label on or off. According to Intercom, it is on by default only for workspaces that were already running Fin in the Messenger before mid-2025, and off by default for everyone who deployed Fin later. Intercom also says its customers are responsible for making sure end users know they are interacting with AI.
So if you set up Fin recently, do not assume you are already covered. Check the setting.
Who is legally responsible? Strictly, Article 50(1) puts the duty on the provider - Intercom, as the maker of Fin. But Intercom's documentation places responsibility on its customers for making sure end users know, and the label and intro message are controlled by you. So in practice, this sits on your desk.
Article 50(5) adds that the information must be clear and distinguishable, given at the latest at the first interaction, and meet the applicable accessibility requirements. In practice: use real text, not only an icon, with enough contrast to read.
Where to put the notice in Intercom
Intercom gives you two built-in tools — a label and an intro message. Use at least one, ideally both. The kit's snippet is optional for Intercom.
1. Turn on the AI Agent label
Find the toggle "Show AI Agent label in Messenger" in Fin's identity settings. Intercom's help centre describes the path as Settings > AI & Automation > Fin AI Agent (under Fin's identity); the Fin help centre describes it as Settings > Reply Settings > Identity. Both lead to the same switch. When on, every Fin message shows an "AI Agent" label after Fin's name.
2. Check Fin's intro message
When you deploy Fin in the Messenger you choose the message Fin opens every conversation with. Intercom's newer default says the visitor is speaking with an AI agent. If you edited it, or you set Fin up before that default existed, make sure your version still says it is AI. Intercom specifically advises this if the label is turned off.
3. Fin's name and avatar
If you renamed Fin to a human-sounding name, keep the label on or say "AI" in the intro. A friendly name is fine; a name plus a photo plus no label is where visitors get misled.
4. Several brands
Workspaces with multiple brands can give Fin a different identity per brand. Check the label and intro for each brand, not just the default.
5. The HTML snippet (optional)
If you want a notice visible before the Messenger opens, paste the kit's snippet in your site footer.
Intercom's two help centres currently describe different menu paths to the same toggle. If neither matches your screen, search Intercom's help for "AI Agent disclosure".
Example notice line
Hi, you're chatting with our AI agent. You can ask for the team at any time.
Use your own wording if you prefer, but keep the words "AI" and a plain statement that it is not a person. The exact wording isn't set by the European Commission — it just needs to be clear.
What this page is not
We are not a partner of Intercom and are not affiliated with or endorsed by them. This page covers Article 50(1) chatbot transparency only — not the rest of the EU AI Act, such as high-risk systems, general-purpose AI models or AI-generated content labelling, and not GDPR. It is general information, not legal advice. If your AI use goes beyond a customer chat, get proper legal advice.
Sources
Maintained by Penn Ltd. Last checked against Intercom's help centre and the regulation text in September 2026. General information, not legal advice.