← Free risk checker
Home → Chatbot Compliance → Intercom EU AI Act
LIMITED RISK — Article 50 transparency required
Deadline: 2 August 2026. Article 50 has NOT been deferred by the Digital Omnibus. You need a disclosure on your Intercom chatbot before this date. Fine for non-compliance: up to €15M or 3% of turnover.

Intercom EU AI Act Compliance: What You Must Add by 2 August 2026

Using Intercom Fin AI or automated chat? Article 50 of the EU AI Act requires a clear disclosure that users are talking to an AI — not a human. Here is exactly what to add and how.

Good news: Intercom is limited risk — not high-risk. You do not need Annex IV documentation or conformity assessment. You need one thing: a clear disclosure that the user is talking to an AI. It takes one hour to add.

What you need to do — step by step

  1. Add a disclosure message as the first message in your Intercom bot flow, before any user input
  2. Make the disclosure clear and prominent — not in a tooltip or help article
  3. Ensure Intercom's bot name reflects its AI nature (e.g. "Fin AI", "Support Bot") — not a human-sounding name with no indication of automation
  4. Update your privacy policy to mention that an AI chatbot is used for customer interactions
  5. If Fin generates written content users will publish or share, mark it as AI-generated (Article 50(2))
  6. Train support staff on the AI literacy requirement — they should understand what Fin can and cannot do (Article 4)

Ready-to-use disclosure text for Intercom

Copy one of these into your Intercom bot's opening message:

Hi there! I'm Fin, an AI assistant. I'm here to help with common questions — type agent at any time if you'd prefer to speak with a person.
Welcome! You're chatting with an automated AI assistant. A human support agent is available if needed — just ask.
Hello! This is an AI-powered chat. I can answer most questions instantly. Type "human" anytime to connect with our team.

Need this in French, German, Spanish, Dutch, Polish, Italian?

The Chatbot Compliance Pack includes 7 language variants, a T&C clause template, a privacy policy AI section, and the Article 50(2) machine-readable marking guide.

Get Chatbot Compliance Pack — €49 →
Or use the free disclosure generator →

Intercom and the EU AI Act: full analysis

Why Intercom Fin is limited risk, not high-risk

Article 50 of the EU AI Act applies to "AI systems intended to interact directly with natural persons." Intercom Fin clearly qualifies. However, the limited risk classification (rather than high-risk) applies because a standard customer support chatbot does not make consequential decisions about people's rights, safety, or fundamental interests.

Intercom Fin becomes high-risk only if you configure it to assist with employment decisions, credit assessments, benefit eligibility determinations, or similar high-stakes functions listed in Annex III. Review your Intercom workflows — if any bot flow collects data used for those purposes, that workflow escalates to high-risk even if the rest of your Intercom setup is limited risk.

Where exactly to add the disclosure in Intercom

The most reliable approach is a bot step at the very start of your Intercom workflow — before any qualifying question. In Intercom's visual bot builder:

Alternatively, set the chatbot's display name to include "AI" or "Bot" — this provides partial disclosure, though the regulation requires an explicit statement, not just a name hint.

The 2 August 2026 deadline is firm

High-risk obligations may be deferred to December 2027 if the Digital Omnibus is formally adopted. Article 50 transparency is not in that deferral. It applies from 2 August 2026 for all chatbots that interact with users in the EU — regardless of where your company is based.

If your Intercom instance serves EU users (you have customers in any EU member state), you need to add the disclosure before 2 August 2026.

Frequently asked questions

We use Intercom but only for internal employee help desk (not customers). Does Article 50 apply?
Yes. Article 50(1) applies to AI systems that "interact directly with natural persons" — this includes internal employees using a help desk bot, not just external customers. You still need a disclosure at the start of interactions, though the practical stakes and enforcement priority are lower for internal tools.
Our Intercom bot sometimes hands off to a human agent. Does the disclosure need to mention this?
The disclosure must be clear that the user is talking to an AI. Mentioning that a human is available is good practice and helps with user experience, but it is not legally required by Article 50. Including it (e.g. "type 'agent' to speak with a person") is strongly recommended as it reduces user frustration and demonstrates good faith compliance.
Is Intercom itself responsible for compliance, or are we?
Both have roles. Intercom as the provider of the AI system has obligations around documentation and transparency. You as the business deploying Intercom on your website are the deployer — and the Article 50 disclosure obligation falls on you as the entity that "puts the system into service" for your users. Intercom providing a compliant product does not substitute for you adding the disclosure to your bot.
We are a US company with some EU customers. Does this apply to us?
Yes. If your Intercom chatbot interacts with users located in EU member states — even if your company is based in the US — Article 50 applies. The EU AI Act has extraterritorial reach: it applies wherever users in the EU are affected, not where the company is incorporated.