Compliance Tools
The EU AI Act (Regulation 2024/1689) classifies AI systems into risk tiers. Your obligations depend on which tier your system falls into. This wizard guides you through the classification in under 5 minutes.
Risk tiers
| Tier | What it means | Example |
|---|---|---|
| Prohibited | Banned outright | Social scoring by governments, real-time mass biometric surveillance |
| High-Risk (Annex I) | Safety component of a regulated product | AI in medical devices, vehicles, lifts |
| High-Risk (Annex III) | Sensitive use case | AI in hiring, education, law enforcement, border control |
| GPAI with Systemic Risk | Foundation model above compute threshold | Models trained with ≥10²⁵ FLOPs |
| GPAI Standard | Foundation model below threshold | Most commercial LLMs used as components |
| Limited Risk | Interacts with people | Chatbots, deepfake generators (must disclose) |
| Minimal Risk | Everything else | Internal analytics, recommendation engines |
How to use the wizard
Navigate to Compliance Command Center → AI Act tab.
- Name your AI system — enter a descriptive name (e.g., "Customer Support Bot" or "Hiring Screening Tool")
- Answer the questions — one question per screen. Each has a "Why does this matter?" explanation you can expand
- Watch the classification badge — it updates live on the right as you answer
- Review the decision path — when complete, you'll see exactly which questions led to your classification
- Save — the classification, decision path, and a formatted memo are saved to your compliance record
The classification is free
The EU AI Act Risk Classifier is available on the free tier. No license required. It's the natural starting point for understanding your compliance obligations.
What the memo contains
The classification memo includes:
- Your AI system name and classification date
- The risk tier result with applicable EU AI Act articles
- A decision path table showing every question, your answer, and the rule reference
- A disclaimer recommending legal review
Frequently asked questions
How long does classifying a system under the EU AI Act take?
The wizard guides you through classification in under 5 minutes, across five steps: name the system, answer questions, watch the classification update, review the decision path, and save the results.
What are the possible risk tiers?
Prohibited, High-Risk (Annex I), High-Risk (Annex III), GPAI with Systemic Risk, GPAI Standard, Limited Risk, and Minimal Risk.
Do we need a paid licence to use the classifier?
No. The EU AI Act Risk Classifier is available on the free tier and no licence is required.
What does the classification memo actually contain?
System name, date, tier result, applicable articles, a decision path table, and a legal review disclaimer.
Glossary
| Term | Definition |
|---|---|
| EU AI Act | Regulation (EU) 2024/1689 — the European Union's comprehensive AI regulation |
| Annex I | List of EU product safety legislation — if your AI is a safety component of these products, it's high-risk |
| Annex III | List of high-risk AI use cases (employment, education, law enforcement, etc.) |
| GPAI | General-Purpose AI — foundation models like Claude, GPT, Gemini, Llama |
| Notified Body | An EU-accredited organisation that performs mandatory third-party assessments for certain high-risk systems |
| Art.50 | The transparency article — chatbots must disclose they're AI, deepfakes must be labelled |