Compliance with the EU AI Act – Blip | Blip Help

To comply with the EU AI Act (the European Union's Artificial Intelligence Act), Blip has natively and automatically implemented the display of transparency messages in conversation flows conducted by AI.

This feature ensures that the end user is clearly and transparently informed whenever they are interacting with an Artificial Intelligence or transitioning to human support.

How does activation work?

To make it easier for your company to comply with the legislation, the feature has been enabled directly on Blip's infrastructure for all accounts and smart contacts operating in the European Union.

This means the feature:

Does not require manual configuration: No action is needed in the portal.
Does not require republishing the flow: The behavior is applied automatically in the background.
Does not block your usage: Activation happens transparently, without interrupting your operations.

Display rules and standardized messages

In this first version, transparency messages are fixed and standardized. The use of non-editable texts ensures the legal security of your operation, guaranteeing that the notice fully meets the legal transparency requirements.

The messages are displayed immediately before the start of a conversation within an AI context, ensuring technical predictability and clarity for the user (see the following image)..

Message mapping in the flow

At this initial stage, the messages are written in English, but they will later be adapted into other languages.

Moment in the conversation	Display rule	Message shown to the user
Start of conversation with AI	Displayed before the AI's first generated response.	ⓘ Before we begin, it's important for you to know that you're talking to an artificial intelligence.
AI resumption (after a deterministic block)	Displayed when the conversation leaves a fixed/scripted block and returns to an AI agent.	ⓘ It's important to remember that your conversation is being conducted by an artificial intelligence.
Transfer to human support	Displayed before connecting the user to a human agent.	ⓘ From now on, your conversation will be conducted by a human agent.
Return from human support to AI	Displayed when human support ends and the flow returns to AI.	ⓘ From here on, your conversation is once again being conducted by an artificial intelligence.
Switching between AI agents	Not repeated. The AI context is maintained without unnecessary interruptions.	(No message is displayed)

How to test the behavior in Blip Studio

Although activation is automatic in production, you can validate the visual experience and response time of your smart contact:

Access Blip Studio.
Open your smart contact's flow.
Start Test mode.
Interact with the bot in scenarios involving the start of a conversation, transition to a human, and return to AI.

The transparency messages will be displayed at exactly the same moment they will appear to your end user in production.

Frequently Asked Questions (FAQ)

Can I customize the text of the transparency messages?

In this first release focused on the European Union, the messages cannot be edited or customized. This ensures the content is strictly aligned with the EU AI Act's legal requirements. The ability to customize is planned for future updates.

Do I need to publish a new version of my flow for the notices to work?

No. The feature has been enabled on the cluster's backend. Any interaction from the activation date onward will automatically display the notices.

Do the messages consume my AI or affect response time?

No. The transparency messages are triggered by Blip's platform infrastructure before the AI agent's processing, and do not consume additional tokens from your language model (LLM).

Need more help? Explore our content at Blip Academy or Blip Community, watch tutorials on our YouTube channel, or get your questions answered in our support channel 😃