The EU AI Act is long. The part that applies to you is usually short, once you know which systems you have and what role you play.
Where the Act starts
The regulation sorts AI systems by risk: prohibited practices, high-risk systems, systems with transparency duties, and the rest. It also separates providers, who place a system on the market, from deployers, who use it. Your obligations depend on both. They phase in over several years; some already apply, including the prohibitions and AI literacy.
Why companies get stuck
Nobody has a complete list of AI systems, including the ones inside vendor tools. Classification needs someone who understands both the law and the system. And the output is often a legal memo that nobody in engineering can turn into work. Meanwhile sales receives questionnaires asking for exactly this documentation.
Engineering-grade, not legalese.
How we approach it
Inventory first, including vendor integrations. Then classification per system, with the reasoning written down and reviewed with your legal counsel. A gap analysis against the obligations that apply, ranked by effort. And a dated roadmap with owners, templates and an evidence checklist, handed over in a workshop with legal and engineering together.
Not legal advice. We work alongside your legal counsel.