AI risk classification
Prohibited, high-risk, limited or minimal risk, with the Annex III category and your role as provider or deployer.
Find out which risk tier your AI system falls into, where it falls short of the EU AI Act, how large your fine exposure is, and what to fix first. Includes a GDPR and data-sovereignty review.
Software, not a law firm. Reports are an automated, informational assessment and are not legal advice. For binding decisions, consult a qualified lawyer.
Two agents analyse your system in parallel, two turn the findings into a board briefing and a costed plan. Fines, deadlines and the overall grade are calculated from the regulation's text, not guessed.
Classifies the risk tier and checks Articles 5, 6 and 9–15.
Reviews GDPR, automated decisions and data transfers.
Turns findings into a plain-language board briefing.
Builds a phased, costed plan that closes every gap.
Describe your AI system or upload one PDF. To scan many documents together, create a workspace.
Automated assessment for information only. It is not legal advice and does not replace a conformity assessment or a lawyer's review.
Every finding cites the article it comes from and the evidence it is based on, so your legal and engineering teams can check it.
Prohibited, high-risk, limited or minimal risk, with the Annex III category and your role as provider or deployer.
Maximum penalties under AI Act Article 99 and GDPR Article 83, from your turnover and SME status.
Risk management, data governance, documentation, logging, transparency, oversight, accuracy and security.
The Article 113 dates relevant to your risk tier, and which have already passed.
Automated decisions, DPIA need, processor agreements and transfers, including US CLOUD Act exposure.
Phased actions, each linked to the gaps it closes, with effort in days and a cost estimate your board can approve.
A short overview of the risk-based approach of Regulation (EU) 2024/1689, the key dates and the penalties.
| Risk tier | Examples | What it means |
|---|---|---|
| Unacceptable risk | Social scoring, manipulative techniques, emotion recognition at work or school, untargeted facial-image scraping | Prohibited under Article 5 since 2 February 2025. |
| High risk | CV screening and hiring, credit scoring, insurance pricing, education and exam scoring, biometrics, critical infrastructure | Risk management, data governance, technical documentation, logging, transparency, human oversight, accuracy and cybersecurity (Articles 9–15), plus conformity assessment and registration. |
| Limited risk | Chatbots, AI-generated or manipulated content (deepfakes) | Transparency duties under Article 50: tell people they are interacting with AI and label synthetic content. |
| Minimal risk | Spam filters, recommendation features, most internal tools | No specific obligations beyond AI literacy (Article 4); voluntary codes of conduct. |
The AI Act applies to providers that develop an AI system or place it on the EU market, and to deployers that use AI systems in their work, as well as importers and distributors. It also applies to companies outside the EU when their AI system's output is used in the EU. Most obligations fall on providers of high-risk AI systems, but deployers have their own duties, such as human oversight, monitoring and, for some, a fundamental rights impact assessment (Article 27).
The EU has proposed changes to some high-risk dates. Always check the current official text on EUR-Lex before planning.
Article 99 sets maximum fines of €35 million or 7% of worldwide annual turnover for prohibited practices, €15 million or 3% for breaches of most other obligations, and €7.5 million or 1% for supplying incorrect information. The higher amount applies to large companies; for SMEs and start-ups, the lower one does. GDPR fines of up to €20 million or 4% can apply on top when personal data is involved.
Start with an inventory of the AI systems you build or use, classify each one, and run a gap analysis for anything high-risk. Then prioritise: stop prohibited uses immediately, put human oversight and logging in place, document training data and testing, and prepare technical documentation for the conformity assessment. AI Act Compliance automates the classification and gap analysis so your team and your lawyers can focus on the decisions.
You're sharing sensitive material. Here is exactly how it is handled.
Only a fingerprint of your description and the resulting report are kept. The text itself is discarded after the scan.
Every query is scoped to your workspace. Passwords are hashed with scrypt, sessions use secure HttpOnly cookies, and every action is in your audit log.
Remove single documents, whole projects or individual reports whenever you want. Deleted files are erased from storage.
Uploaded documents are encrypted at rest with AES-256-GCM.
No. AI Act Compliance is software, not a law firm, and nothing it produces is legal advice. The report is an automated, informational assessment that helps you prepare: it shows where your AI system likely stands under the EU AI Act and GDPR and what to fix first. For binding decisions, conformity assessments or regulator correspondence, work with a qualified lawyer or notified body.
The EU AI Act (Regulation (EU) 2024/1689) is the European Union's law on artificial intelligence. It entered into force on 1 August 2024 and sorts AI systems into four risk tiers: prohibited (unacceptable risk), high risk, limited risk with transparency duties, and minimal risk. Obligations depend on the tier and on your role as provider, deployer, importer or distributor.
An AI system is high-risk if it is a safety component of a product covered by EU harmonisation law (Annex I) or if it is used in an Annex III area: biometrics, critical infrastructure, education, employment and HR, access to essential services such as credit scoring and insurance, law enforcement, migration, and justice. Article 6(3) allows some exceptions. The free scan classifies your system and explains why.
Under Article 99, prohibited practices can be fined up to €35 million or 7% of worldwide annual turnover, whichever is higher. Breaches of most other obligations, including high-risk requirements, can be fined up to €15 million or 3%. Supplying incorrect information to authorities can cost up to €7.5 million or 1%. For SMEs and start-ups the lower of the two amounts applies.
Prohibited practices and AI literacy duties have applied since 2 February 2025. General-purpose AI model obligations have applied since 2 August 2025. Under Article 113, most remaining obligations, including Annex III high-risk requirements, apply from 2 August 2026, and Annex I product rules from 2 August 2027. The EU has proposed adjusting some high-risk dates, so check the current official text before you plan.
No. GDPR still applies whenever your AI system processes personal data. That is why every scan also checks GDPR topics that matter for AI: lawful basis, automated decision-making (Article 22), processor agreements (Article 28), DPIAs (Article 35) and international data transfers (Chapter V).
For the free scan, your description isn't stored; only a fingerprint and the resulting report are kept. In a workspace, the documents you upload are stored in your own isolated workspace so you can rescan them, and you can delete them at any time. The analysis is performed by Anthropic's Claude models through their commercial API.
In a workspace you can upload many files at once: PDF, Word (.docx), Excel (.xlsx), CSV, text and Markdown, JSON, YAML, HTML, and images such as architecture diagrams or screenshots. Large document sets are read document by document before the assessment.
Run a free AI Act risk assessment now, or create a workspace to scan your full documentation set.