Does your tool use inference (learning from data to generate predictions, content, or decisions) or just follow fixed rules?

Fixed rules only — no machine learning or inference (e.g., RPA, spreadsheets, if-then logic) General-purpose AI tools (ChatGPT, Copilot, Gemini) Custom ML models for business decisions AI for hiring, credit, healthcare, or law enforcement

Does your AI system make or influence decisions about people?

No — it's used for internal analytics only It suggests but humans always decide It directly influences outcomes for individuals It makes autonomous decisions about people

In which sector does your AI operate?

Marketing, content, or general productivity Finance, insurance, or e-commerce Healthcare, education, or employment Law enforcement, immigration, or critical infrastructure

Does your AI process biometric data (face, voice, fingerprint)?

No biometric data at all Voice for transcription only Face/voice for identification or verification Real-time biometric identification in public

How many people are affected by your AI system?

Internal team only (< 50 people) Hundreds of customers or employees Thousands of users Tens of thousands or more

Do you have documentation for your AI system?

Comprehensive technical docs and risk assessments Some documentation, not formal Minimal — we know how it works internally No documentation

Is there human oversight in your AI's decision process?

Yes — humans review every AI output Humans can override but don't always review Limited oversight on edge cases only Fully automated, no human in the loop

Do you inform users they are interacting with AI?

Yes, clearly and prominently Mentioned in terms of service Not explicitly Users believe they interact with humans

How do you handle AI system errors or failures?

Formal incident response + logging + audit trail We log errors and fix them when found Ad-hoc fixes, no formal process No error monitoring in place

Does your company operate in or serve customers in the EU?

No EU presence or customers Some EU customers but not primary market Significant EU customer base Headquartered in the EU

Do you use or integrate General-Purpose AI models (e.g., LLM APIs)? (GPAI context)

No Yes, for internal drafting only Yes, in customer-facing workflows Yes, in high-impact decision workflows

Are AI transparency notices implemented for users? (Article 50)

Yes, clearly visible in-product Partially implemented Only in legal terms No transparency notices

If a use case is high-risk, are conformity and technical documentation workflows defined? (Art. 11/assessment)

Yes, fully defined In progress Not formalized No process

Do you run post-market monitoring and periodic reassessment for deployed AI?

Yes, with scheduled reviews Some monitoring, irregular Ad-hoc only No monitoring

Have relevant staff received AI literacy training? (Article 4)

Yes, role-based training completed Basic awareness only Planned but not delivered No training

Get your personalized risk report

Enter your details to receive your AI Act risk classification and a compliance checklist.

Work email *

Name

Company

Company size

EU AI Act Risk Assessment