Practical writing on AI governance, EU AI Act compliance, and professional systems.
The 14 questions a customer or auditor can ask a high-risk AI provider, mapped to articles of the EU AI Act and to the evidence worth keeping.
A practical framework for classifying AI systems by risk tier and identifying which obligations apply now versus later, under the 2026 regulatory timeline.
The three specific triggers under Article 25 that can reclassify a deployer into a provider overnight — rebranding, substantial modification, and purpose change — and why the Digital Omnibus didn't make this easier, only later.
Plain-language definitions of the terms that matter most when classifying and governing AI systems — provider vs. deployer, high-risk classification, GPAI, and more.
More guides are added regularly. For the underlying tools, see Business Frameworks →