As of August 2, contact centers must ensure that their chatbots and voicebots identify themselves as AI. This is one of the obligations arising from the third phase of the EU AI Act, in which transparency plays a central role.

Anyone using AI in customer contact must clearly inform the customer of this in advance. In addition, organizations must be able to demonstrate and document how they use AI for applications such as selection or assessment. The European Commission has published additional guidelines and ready-made icons for this purpose.

Response suggestions that an employee reads first and then sends fall outside the scope of the AI ​​Act, but a fully automated email drafted without human intervention may well fall under the AI ​​Act. In addition, Dutch regulators are beginning to enforce AI literacy, a requirement since August 2025. Contact center managers must be able to demonstrate, by role, who has been trained and in what way. After all, an agent using an agent-assist tool requires different knowledge than a WFM specialist using AI for forecasting.

Text continues below the image

AI Act: what contact centers must have in place by August 2, 2026

Enforcement

As of August 2, national supervisory authorities are expected to begin enforcement. In the Netherlands, the Dutch Data Protection Authority and the National Inspectorate for Digital Infrastructure will assume a coordinating role, with the AFM and DNB, among others, additionally responsible for the financial sector.

Compliance and a successful AI implementation require a good preparationConsider an internal AI register in which you record the user, purpose, data, risk, and ownership for each application. Such a register is not mandatory, but it can be a useful tool for demonstrating compliance. The Algorithm Framework now also offers practical solutions. tools, such as guidance on determining whether something is an AI system and fact sheets on the AI ​​Regulation and AI literacy.

(Ziptone/editors)

Interesting article? Share it!