The EU AI Act without the fussSince July 2026 there's a revised timetable, and Germany has a new supervisory authority. What companies using AI need to do now, and what can wait until 2027.AI governance & regulationFurther reading
Topic · AI governance & regulation
Good rules make you faster, not slower.
A register of the AI systems in use and clear responsibilities make approvals predictable. Then a new initiative doesn't have to start from scratch.
The EU AI Act classifies AI systems by risk and attaches obligations for providers and deployers. For most companies the work begins with the question of which systems they use and in what role, including in software they've bought.
Building on that, responsibilities, review steps and documentation can be designed to fit the existing controls for procurement, data protection and information security. A management system under ISO/IEC 42001 can be built on top.
Context
What’s changing and how we see it
Where things stand in November 2026
Regulation (EU) 2024/1689 applies in stages. Since 2 February 2025 the prohibitions and the obligation to take AI literacy measures have applied, and since 2 August 2025 the rules for general-purpose AI models. Since 2 August 2026 the regulation has been largely applicable, including the transparency obligations under Art. 50. The amending Regulation (EU) 2026/1744, published in the Official Journal of the EU on 24 July 2026, moves the obligations for high-risk systems under Annex III, such as recruitment or consumer credit scoring, to 2 December 2027, and for AI in regulated products under Annex I to 2 August 2028.
What applies in Germany
Since 29 July 2026 the German Act on Market Surveillance and Innovation Promotion for Artificial Intelligence (KI-MIG, Federal Law Gazette 2026 I No. 223) has been in force. The Federal Network Agency (Bundesnetzagentur) is the central market surveillance authority unless the act provides otherwise; BaFin is responsible for financial services.
Our view
Governance acts as a brake when it only kicks in after the pilot. That's why we start with the register, roles and review steps, so every new initiative can be checked against known rules. The legal assessment of individual systems is a matter for your legal department or law firm; we prepare the information they need for it.
What matters
Ground rules
Create an overview
A register of the AI systems in use, with role, purpose and risk class.
Clarify accountability
For every system it's clear who approves, who monitors and who documents.
Keep records
Documentation that stands up to audits without paralysing day-to-day work.
Key questions
What you should clarify
Which AI systems do we use, and in what role?
An inventory that also covers AI features in purchased software will show you. For each system we record purpose, role and risk class.
When do we become a provider rather than a deployer?
When you market a high-risk system under your own name, substantially modify it or repurpose a system so that it becomes high-risk (Art. 25). You then take on a provider's obligations.
Who approves new AI applications?
A named function with clear criteria, built into your existing review processes. A dedicated committee pays off mainly if you operate high-risk systems.
Is one AI policy for all staff enough?
It's a good start, because it sets out which tools may be used with which data. Higher-risk systems also need review steps and documentation for each system.
How we help
Relevant capabilities
- Strategy & Consulting · DecideDeciding where AI pays off, and where it does not.
- AI strategy and roadmap
- Use-case portfolio and business cases
- Operating model
- AI governance
- Academy · EnableAnyone deciding on AI should understand it.
- Executive Education
- Enablement programmes
- Keynotes
Insights
Related articles
The EU AI Act without the fussSince July 2026 there's a revised timetable, and Germany has a new supervisory authority. What companies using AI need to do now, and what can wait until 2027.AI governance & regulationFurther reading
Twelve questions for boards and managing directorsSenior leadership doesn't need to know how a language model works. It should ask the right questions and be able to tell good answers from weak ones.AI & workFurther reading
Agents are employees without a contractAgents open cases, change data and send messages. Four levels of autonomy, a probationary period with fixed metrics and a one-page profile bring order to their use.AI agentsFurther reading