Frameworks
EU Artificial Intelligence Act
The world's first comprehensive AI law — phasing in now, reaching far beyond Europe.
Book a demoThe EU AI Act regulates AI the way GDPR regulated data: by risk tier, with obligations that follow the market rather than the headquarters. If you sell AI-powered products into Europe or your systems touch EU users, its high-risk requirements — risk management, documentation, human oversight, post-market monitoring — are your requirements.
Risk tiers decide everything
The Act bans a short list of practices outright, imposes transparency duties on general-purpose models, and reserves its heavy machinery for 'high-risk' systems: AI in hiring, credit, insurance pricing, education, essential services, and critical infrastructure. Classification is the pivotal exercise — many SaaS companies discover a single HR-screening or credit-adjacent feature puts them in scope.
How to prepare without panic
The preparation order that works: inventory every AI system you build, buy, or embed; classify each against the Act's tiers with documented reasoning; stand up risk management and human oversight for anything plausibly high-risk; and wire post-market monitoring so AI incidents actually reach accountable owners. If that sounds like ISO 42001, it is — the standard operationalizes most of the Act, which is why smart teams run one program for both.
How Compriska helps
Compriska's AI Governance module was designed around this overlap: one AI inventory, risk assessments aligned to both the Act and ISO 42001, approval workflows with human oversight built in, and monitoring that feeds incidents back into the register — one program, two regulators satisfied.
Frequently asked questions
Does the EU AI Act apply to US companies?
Yes — like GDPR, it follows the market. Providers placing AI systems on the EU market and deployers using AI within the EU are in scope regardless of where they're incorporated.
When do the high-risk obligations take effect?
The Act phases in through 2027: prohibited practices and AI-literacy duties applied first, general-purpose model obligations followed in 2025, and the full high-risk regime lands through 2026–2027. The runway is months, not years.
Other frameworks
Automate EU AI Act with Compriska
See how cross-mapped controls, continuous evidence, and AI workflows change the cost of compliance.
Book a demo