EU AI Act - Compliance and Risk Classification

EU AI Act risk classification, gap analysis, and a phased compliance roadmap. Dates, bans, and what a certificate does not prove.


What we offer

The EU AI Act (Regulation 2024/1689) is the world’s first comprehensive AI regulation. It introduces a risk-based classification system for AI systems, with requirements ranging from transparency obligations to full conformity assessments.

We help you cut through the regulatory complexity and build a practical compliance roadmap. Our EU AI Act consulting covers:

Our approach

The EU AI Act is complex, but compliance doesn’t have to be overwhelming. We break it down into manageable steps:

  1. AI system inventory - We help you identify and catalog every AI system in your organization, including those you might not think of as “AI.”
  2. Risk classification - We determine the risk level for each system based on the regulation’s criteria. Many systems are minimal risk. That tier has no conformity assessment and no Article 50 transparency duty. Article 4 still requires providers and deployers to take measures to support AI literacy, without requiring any person to reach a set level. Transparency is the limited-risk tier, including Article 50. It isn’t a minimal-risk duty.
  3. Gap analysis - For systems that need attention, we identify exactly what’s missing and what needs to change.
  4. Compliance roadmap - We create a prioritized plan that aligns with the regulation’s phased enforcement timeline.
  5. Implementation support - We work with your team to implement the required changes, from documentation to technical controls.

Why choose us

The EU AI Act doesn’t exist in isolation. It intersects with GDPR, the NIS2 Directive, and sector-specific regulations. Our experience with EU regulatory frameworks means we understand how these pieces fit together.

We also know ISO 42001. If you’re pursuing both ISO 42001 certification and EU AI Act compliance, there’s significant overlap. We’ll help you build a unified approach that covers the overlap without duplicating work. Certification isn’t a legal finding of compliance. The date-by-date reading is in the EU AI Act timeline.

And honestly? Most of your AI systems probably fall under minimal or limited risk. We’ll tell you that upfront instead of inflating the scope to justify higher fees.

EU AI Act timelines

The EU AI Act entered into force in August 2024. Regulation (EU) 2026/1744 (the Digital Omnibus on AI, in force since 27 July 2026) moved the high-risk dates:

Next step

Not sure how the EU AI Act applies to your AI systems? Let’s have a conversation. We’ll help you understand your obligations - and if the honest answer is “you’re mostly fine,” we’ll tell you that too.

Let’s talk - no commitment, just clarity on where you stand.


Frequently Asked Questions

Does the EU AI Act apply to companies outside the EU?
Yes, in the cases Article 2 sets. It covers providers who place an AI system on the market or put it into service in the Union, and providers who place a general-purpose AI model on the Union market. It covers deployers established or located in the Union, and providers or deployers outside the Union where the output is used in the Union. Importers, distributors, and product manufacturers have duties too.
How do I know if our AI systems are high-risk?
Article 6 has two paths. Article 6(1) and Annex I cover an AI system that is a safety component of a regulated product, or is itself that product. Article 6(2) and Annex III list use cases such as employment, credit scoring, and law enforcement. That list is longer than the examples. We'll help you classify every system in your inventory.
What are the penalties for non-compliance?
Fines are up to 35 million EUR or 7% of worldwide annual turnover for prohibited practices, and up to 15 million EUR or 3%, or 7.5 million EUR or 1%, otherwise. For an undertaking, each ceiling is the higher of the two. Article 99(6) is up to the lower figure for SMEs, including start-ups. Article 99(6a) is up to the lower figure for small mid-caps on the 3% and 1% tiers only. The prohibited-practice ceiling stays the higher of the two for a small mid-cap.
What about a high-risk system already on the market?
Article 111(2) applies Chapter III to a high-risk system placed on the market or put into service before the relevant date only if the design changes significantly after it. Recital 39 of Regulation (EU) 2026/1744 lets further units of the same type and model continue while that design is unchanged. Systems intended for public authorities have until 2 August 2030.
When do we need to be compliant?
Prohibited practices have applied since February 2025, and GPAI obligations since August 2025. Article 50 applies from 2 August 2026. From 2 December 2026 further bans apply, as limited by Article 5(1a) and (1b): non-consensual intimate material (Article 5(1)(ba)), and what recital 11 calls child sexual abuse material (Article 5(1)(bb)). The operative text is child pornography and pornographic performance under Directive 2011/93/EU Article 2(c) and (e). Article 111(4) sets Article 50(2) marking by that December date only for synthetic audio, image, video or text systems, including general-purpose AI, placed on the market before 2 August 2026. Chapter III, Sections 1, 2 and 3, except Article 6(5), apply from 2 December 2027 (Annex III) and 2 August 2028 (Annex I).
Can we handle EU AI Act compliance and ISO 42001 together?
Yes, and we recommend it. There's significant overlap. ISO 42001's AI risk assessment, documentation, and monitoring requirements map to many EU AI Act obligations. A unified approach saves time and reduces duplication. Certification isn't a legal finding of compliance.