What are the EU AI Act risk tiers?

The EU AI Act defines four risk tiers: unacceptable (prohibited), high (strictest obligations), limited (transparency obligations), and minimal (no mandatory obligations). The classifier assigns your system to one of these.

When does the EU AI Act apply?

Current EU guidance lists 2 August 2026 for most AI Act rules, with high-risk systems embedded in regulated products following on 2 August 2027.

Is the classification legal advice?

No. GeraCompliance provides structured compliance guidance, not legal advice or a notified-body assessment. Results should be reviewed by qualified counsel before external reliance.

EU AI Act enforcement is now in effect (since 2 August 2026)Learn more

Step 1 of 3

AI System Risk Classification

Describe your AI system below. This free check gives a preliminary tier from the words in your description. It is a keyword based screen, not a full assessment, and it is not legal advice.

Free sample evidence pack

See what the evidence behind a classification looks like. The free sample pack holds seven anonymised PDF excerpts from a paid GeraCompliance sprint, including a risk classification memo and an Annex IV technical documentation skeleton. Free to download, no email signup. After you classify a system above, you can print or save your own result as a PDF from the result panel.

Browse the free samples Risk classification memo (PDF)

Full AI Act Assessment Report

£349

Your free classification written up in full: the complete obligations register for your risk tier, an Annex IV technical documentation skeleton for your system, and a 30 day action plan, generated from the details you classified. Classify your system above first, free, and the full offer appears with your result — or go straight to the report if you already know your tier.

Get the full report · £349 See a sample report first

Generated on screen as soon as payment completes.