Useful context before an assessment.

Start with the system and its use.

The EU AI Act sets rules according to risk and role. Providers and organisations using a system may have different obligations. Some practices are prohibited; some systems face high-risk or transparency requirements.

Qualification depends on the specific intended purpose. An internal policy, use inventory and documented reviews are governance tools: on their own, they are not universal proof of compliance.

A timeline with exceptions.

The AI Omnibus entered into force on 27 July 2026, changing high-risk timelines and simplifying AI literacy provisions. Transitional rules and the system’s context need to be checked.

European Commission · AI Act · AI Omnibus

Prepare a useful assessment.

Here is what I suggest gathering for an initial diagnostic:

1. Current and planned uses, their purpose and their users.
2. The suppliers, models and integrations involved.
3. The data used and the people who could be affected.
4. Existing rules, approvals and controls.
5. Customer requests, deadlines and unresolved questions.

My role is then to organise this information, examine the scope and recommend priorities. Questions requiring a formal legal opinion are addressed with your advisers.

Sources and scope

This guide provides general context. For a decision, consult the applicable texts and current guidance, then examine your situation.

• European Commission: AI regulatory framework
• European Commission: AI Omnibus enters into force
• AI Act Service Desk