A brand of BritixoShadow AI. Clear accountability.

Guide · reviewed 21 October 2026

A practical starting point for the EU AI Act.

Start with your systems, their intended uses and your role. This guide explains the questions to organise before turning a general regulatory concern into a programme of work. Reviewed 21 October 2026.

By EUAIC editorial team · Updated

Check the activity and role before the checklist.

The Act can apply to providers placing AI on the EU market, EU deployers and some non-EU activities where outputs are used in the EU. Scope and exceptions require a factual review; a company's registered address alone does not settle the question.

For each use case, identify who supplies the system, who operates it and how the output is used. Keep those facts separate from the legal conclusion so a reviewer can test the reasoning and update it when the deployment changes.

Separate the questions that need different answers.

Consider whether the activity is in scope, whether a prohibited practice is involved, whether high-risk criteria apply and whether specific transparency duties are relevant. General-purpose model responsibilities form another part of the framework. Avoid treating these as a single severity score.

The useful output is a set of reasoned decisions and assigned questions. For example, an uncertain supplier role should remain an explicit action rather than being concealed behind an overall readiness percentage.

Use the current application dates.

The AI Act has phased application. The 2026 amendments changed important high-risk milestones. The official Article 113 text is the reference for the applicable provisions and exceptions.

  • General application: 2 August 2026, subject to exceptions.
  • Relevant Chapter III provisions for Annex III high-risk systems: 2 December 2027.
  • Relevant Chapter III provisions for Annex I product-related high-risk systems: 2 August 2028.
  • Earlier provisions and other applicable laws can already create responsibilities; the later high-risk dates are not a general pause on governance.

Turn the review into specific work.

Build an inventory with clear purposes and owners. Record the role and classification questions, identify missing supplier or technical evidence and assign people to resolve them. Define how approval conditions, oversight and changes will be managed after the first review.

Use this guide to prepare a better discussion with your legal, technical and operational advisers. EUAIC can support the records and workflow around that discussion; this general guide is not a legal determination or certification of a system. Keep the official sources and review date with the programme.

Sources & further reading

Useful answers

Before you take the next step.

Practical details to help you discuss your requirements.

Do later high-risk dates mean we can ignore AI governance now?

No. Other AI Act provisions and applicable privacy, employment, product or sector rules may already matter. Check the specific activity and requirements.

Is being outside the EU enough to be out of scope?

No. Certain non-EU activities can be covered. Review where systems are supplied, where outputs are used and the relevant statutory conditions.

What should we do first?

Create a reliable description of the priority systems and uses, identify accountable owners and list the questions that need specialist review.

Start with the right questions

Tell us what you need to achieve.

Share your current setup, priorities and intended outcome. The Britixo team will help define the scope, responsibilities and next steps.

Discuss your requirementsBritixo MyAccountVerify your email. Send your enquiry. Keep your reference.

Explore services and practical guidance.

Search matching runs in your browser. Press Esc to close. Browse every page ↗