Eltrus is monitoring the EU AI Act implementation timeline. The 7 May 2026 Digital Omnibus deferred Annex III high-risk AI obligations to 2 December 2027, but Article 4 (AI literacy) and Article 50 (transparency) enforce on 2 August 2026 regardless. Both apply to Eltrus's service.
Provider and Deployer roles
Eltrus acts as a Deployer of third-party general-purpose AI models (Anthropic Claude, OpenAI GPT) when running AI-assisted customer support inside a Client's helpdesk. The Client may be a co-Deployer depending on how the workflow is configured. The MSA and AI Governance Schedule allocate Provider, Deployer, and downstream-Deployer obligations explicitly per engagement.
Article 50 transparency (effective 2 Aug 2026)
Where the AI is responding to a customer in real time, Client end-customers are informed they are interacting with an AI system, in line with Article 50(1) GDPR-style obligations. The exact transparency language is set in the AI Governance Schedule for each engagement.
Article 22 GDPR automated decision-making
Where the AI materially influences a refund, cancellation, or subscription decision (typically Tier 2 revenue-recovery flows), the workflow is configured with a confidence threshold above which the decision routes to a human agent. The Client owns the policy that defines the threshold.
Article 4 AI literacy (effective 2 Aug 2026)
Eltrus maintains documented AI literacy for personnel deploying or operating AI on Client systems. On request, we provide a short literacy summary for the Client's own personnel using the deployed system.
Model training
Customer data is not used to train any foundation AI model. Anthropic and OpenAI API modes are configured to opt out of training where the providers offer that setting. Zero Data Retention (ZDR) is requested where the Client's risk profile warrants it.