— AI governance practice
kalby
Turning AI regulation into work your teams can schedule — anchored in the EU AI Act, extended to wherever else you operate, and revised whenever the rules move.
scroll ↓ follow the neck
01 The threshold
It starts with which tier you’re in.
Every AI use case sits somewhere on the risk ladder — unacceptable, high-risk, limited-risk, minimal-risk. That single classification decides what you owe, and just as importantly, when.
We map your systems to the right tier first, so the work lands where the obligations actually are — and not where they aren’t.
02 What we do
From board briefings to technical deep-dives.
Gap assessments, AI use policies, risk-classification worksheets, and role-appropriate training — calibrated to the audience, whether that’s an executive team or the engineers shipping the model.
Binding requirements and voluntary standards, always clearly separated. We’ll tell you what the law demands and where a qualified lawyer needs to weigh in.
We also build free tools, because some of this shouldn’t need a consultant. kalby.overview maps which AI Act articles apply to you, by role and risk class, with every date updated for the Digital Omnibus. kalby.check tells you in one click whether your site carries the Article 50 notice, and kalby.disclose adds it in all 24 official EU languages from a single line of HTML. All three are open source and free to use.
03 Approach
Plain language, practical steps.
Advisory, not alarmist. We translate a fast-moving rulebook into clear decisions your teams can act on, grounded in current official sources rather than hearsay.
The goal is a governance posture you can maintain — not a binder that gathers dust after the audit.
say hello.
Tell us about you. We usually reply within a day.
Thanks. Your message is almost on its way — we’ll be in touch.