Help Scout and the EU AI Act: does your chat widget tell people it's AI?
Help Scout's Beacon widget offers AI Answers, an agent that responds using your Docs knowledge base, configured per Beacon at Manage > Beacons > [Beacon] > Edit Beacon > AI Answers, with a choice between "Self Service" mode (AI Answers first, other options after) and "Neutral" mode (AI Answers shown alongside Email/Chat/Docs from the start). Because AI Answers reads and responds in natural language rather than just linking to articles, it's a live AI interaction under Article 50(1), and Self Service mode is specifically the configuration where it's the first, not optional, thing a visitor meets.
The rule itself is short. Article 50(1) of the EU AI Act requires AI systems that interact directly with people to be designed so users know they're dealing with AI, no later than the first interaction, in a clear and distinguishable way. The exception for cases where it's "obvious" is narrow: a natural-language customer-service bot doesn't qualify just because it has a robot icon. A line in your terms of service doesn't satisfy it either. And it's easy to enforce, because a regulator can open your site and start a chat.
Where the disclosure lives in Help Scout
- The Beacon widget launcher and its opening screen, which differs between Self Service and Neutral mode
- AI Answers' presented identity, whatever voice/branding you've set for the agent
- AI Answers' generated response to a visitor's question, before any human sees the conversation
- The path from an unresolved AI Answer to a human channel (Email or Chat), and whether that transition is explicit
Your Help Scout disclosure checklist
- Open a Beacon-enabled page in a private window and ask a question: does anything identify the response as AI-generated before or within the first answer?
- Check your Beacon mode (Self Service vs. Neutral) at Manage > Beacons > [Beacon] > Edit Beacon > AI Answers, Self Service puts the AI response first, which raises the bar for a clear first-interaction disclosure
- Review the voice/brand directives you've set for AI Answers for anything that makes it sound like a specific human teammate rather than an AI system
- Add an explicit AI statement to the AI Answers opening screen or first response template
- Confirm the escalation to Email or Chat clearly signals a change from AI-generated answers to human correspondence
- Screenshot the Beacon's first-interaction state in your configured mode and export the AI Answers settings with a date
Watch out for
AI Answers is framed as a knowledge-base search feature, which can make it feel exempt from "chatbot disclosure" thinking, but it generates natural-language responses conversationally, not just search results, and Self Service mode puts it in front of every visitor before any other option. The product's docs-and-search framing is a reason teams overlook the disclosure question, not a reason the duty doesn't apply.
Common questions
AI Answers just summarizes our help docs, is that really an "AI interaction" under Article 50?
If it's generating conversational natural-language responses to a visitor's specific question, yes, that's the interaction pattern Article 50(1) targets, regardless of how constrained the source material is.
We use Neutral mode, so visitors can skip straight to human chat, does that help?
It gives visitors an obvious alternative, which is good UX, but it doesn't remove the duty for visitors who do engage with AI Answers. Put the disclosure inside the AI Answers flow itself.