Ada and the EU AI Act: does your chat widget tell people it's AI?
Ada is an AI-first support platform — there is no 'human live chat' mode to hide behind: what visitors meet in the widget is the AI Agent by design. That should make Article 50(1) easy, but Ada's own persona tooling cuts the other way: its best-practice docs recommend a human, brand-attached name ("Ada from Acme") and advise avoiding labels like 'Bot'. Follow that advice literally, add a face-style avatar, and you've built a bot that presents as a person — which is precisely when the AI Act's 'obvious from context' escape hatch stops applying and an explicit disclosure has to carry the weight.
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 simply open your site and start a chat.
Where the disclosure lives in Ada
In a typical Ada setup, the surfaces that can carry (or fail to carry) the AI disclosure are:
- The chat button frame (#ada-button-frame) — the entry point, before any conversation exists
- The AI Agent's identity block: the Agent name and avatar configured under Config > AI AGENT > Settings > Preferences
- The opening message of a new conversation — the moment Art. 50(1) cares about
- Human handoff moments — Ada hands to your agents; the participant change should be visible in both directions
Your Ada disclosure checklist
- Open your site in a private window as an EU visitor and start a chat: is it explicit you're talking to AI before or in the very first exchange?
- In the Ada dashboard, go to Config > AI AGENT > Settings > Preferences: check the Agent name — a human-sounding name with no AI qualifier ('Sarah') needs an explicit AI line elsewhere; a name like 'Acme AI Agent' does the disclosure work itself
- Review the avatar choice in the same Preferences tab: a photo-style human face plus a human name is the exact 'non-obvious' combination to avoid
- Put an explicit line in the greeting (e.g. "I'm Acme's AI agent — I can bring in a human anytime") rather than relying on the name alone
- Verify the disclosure survives on mobile and in the compact widget state
- Screenshot the first-interaction state and export your persona settings with a date
Watch out for
Ada's persona guidance optimises for trust and warmth — human name, no 'Bot' labels. Each choice is reasonable growth advice; together they strip the 'obvious from context' exception exactly when your EU traffic needs it. And because persona settings are a dashboard toggle any admin can change, a rebrand or A/B test can silently remove the AI cue your compliance posture depended on — drift the next re-scan should catch.
Common questions
Ada is 'AI-first' — isn't it obvious to users that they're talking to AI?
Not automatically. The Act's exception applies when AI nature is obvious to a reasonably well-informed person from context. A widget named with a human first name, a face-style avatar and natural conversation reads as human chat to many visitors. If you use Ada's humanising persona options, pair them with an explicit AI line at first interaction.
Where do I change what Ada calls itself?
Config > AI AGENT > Settings, Preferences tab: Agent name, company name and avatar live there (Ada's Persona docs cover the same options via API). The greeting text is set in your conversation flows — that's the surface where an explicit AI disclosure belongs.
Check it in one scan.
DisclosureProof opens your site the way a regulator would, triggers your Ada widget, and records whether the AI disclosure actually appears — with timestamped evidence either way.
Scan your site free