The EU AI Act is coming: what is the impact on your SME?
August 2025 - The European Union is gradually introducing stricter rules on artificial intelligence. The AI Act, which has been officially in force since 1 August 2024, will fundamentally change the use of AI within companies. In less than a month, on 2 August 2025, the first major obligations will come into force. For those who are already experimenting with AI in their SME – or who want to do so soon – now is the time to prepare.
The first milestone of the legislation is actually already behind us: since 2 February 2025, AI applications with an unacceptable risk have been banned. These include systems that rank people through “social scoring”, AI that attempts to read emotions at school or in the workplace, or applications that manipulate vulnerable individuals. Such systems are no longer permitted on the European market. For most SMEs, this ban has no direct impact, but it does show how strict and fundamental this legislation is.
Transparency obligations from 2 August
The next important step will follow shortly. From 2 August 2025, transparency obligations will apply to AI that is much more common in practice, such as chatbots, automatic recommendations or systems that generate text or images. Companies will then have to make it clear when customers or users interact with an AI system. General AI models – for example, language models such as GPT – are also covered by the new obligations. Anyone who integrates such technology into their own application must be able to demonstrate how the system works and what risks are associated with it.
For many SMEs, this will be an additional administrative burden, especially if they purchase AI solutions from external suppliers. However, this new law also offers opportunities. The AI Act aims not only to regulate, but also to build trust: customers and partners must be confident that companies are using AI in an ethical and transparent manner. Those who get this right now will have a head start later on.
The rules for so-called high-risk applications, such as AI in HR tools or credit assessment, will follow in August 2026. But preparation is important here too. By mapping out your AI use today, you can avoid unpleasant surprises and build a future-oriented and legally compliant approach step by step.
What should you do today as an SME?
1. Map out your AI usage
Do you use tools or software that incorporate AI? Think of chatbots, HR tools, predictive analytics, customer systems or generative AI (such as automatic text writers). Make a list of which systems you use, what you use them for and who supplied them.
2. Assess the risk
Check whether your applications fall under a prohibited or high-risk category according to the AI Act. For example, do you use AI for recruitment, credit analysis or decisions about access to services? If so, you may be subject to the stricter regime that will come into force in August 2026. However, even low-risk tools such as AI chatbots will be subject to transparency requirements from August 2025.
3. Take the first steps towards compliance
Check with suppliers: does their software comply with the AI Act?
Add a clear notification to AI-driven customer tools (“This is an AI assistant”).
Start with simple documentation: what does the system do, what data does it process, who supervises it?
