Data Protection, AI & Digitalisation

The EU AI Act in 2026: what SMEs need to implement now

The European AI Regulation has been applying in stages since 2024. The Digital Omnibus adopted in summer 2026 postponed individual deadlines – but by no means all of them. For small and medium-sized enterprises, the question is therefore less whether the AI Act affects them than which obligations already apply today.

What it is about

The AI Regulation (Regulation (EU) 2024/1689) governs the use of artificial intelligence according to a risk-based approach: certain practices are prohibited, so-called high-risk AI systems are subject to strict requirements, and other systems are mainly subject to transparency obligations. Most everyday applications in mid-sized companies – drafting texts, translation, research – fall into the lower tiers.

Provider or deployer?

The obligations depend largely on the role. Providers develop an AI system or place it on the market under their own name. Deployers use an AI system under their own authority in a professional context. Most SMEs are deployers: they use tools such as ChatGPT, Copilot or AI features in their software.

Caution is required if a company offers a third-party system under its own brand or substantially modifies it. It may then itself assume the role of provider, with considerably more extensive obligations.

What already applies

  • Since February 2025, certain AI practices have been prohibited, such as manipulative techniques or social scoring.
  • Also since February 2025, there has been an obligation to ensure sufficient AI literacy among the people who use AI systems in the company. The Digital Omnibus adjusted the content of this obligation but did not abolish it.
  • Since 2 August 2026, the transparency obligations have applied: anyone who uses a chatbot on their website, for example, must make it clear that users are interacting with an AI.

What the Digital Omnibus has postponed

With Regulation (EU) 2026/1744, in force since 27 July 2026, the European legislator has postponed the date of application of the obligations for high-risk AI systems: for stand-alone systems under Annex III to 2 December 2027, and for systems embedded in products under Annex I to 2 August 2028.

Important

Annex III includes, among others, AI systems used in recruitment and workforce management and in creditworthiness assessments. Companies using such tools have more time – but preparation should start now.

Five steps for SMEs

  • Create an AI inventory: Which AI systems are used in the company, by whom and for what – including AI features in existing software?
  • Classify the risk: Does an application fall into a high-risk area, such as recruitment? Is the company a deployer or a provider?
  • Issue a usage policy: Which tools are approved, which data may be entered, who checks the results?
  • Ensure AI literacy: Train employees in line with their actual use and document it.
  • Review transparency and contracts: Labelling of chatbots and AI content, data protection and the contracts with AI providers.

The regulation provides for substantial fines for violations. For SMEs, the lower of the amounts provided for applies in each case – noticeable relief, but no reason to postpone compliance.

AI Check under the AI Act

In our module, we record your use of AI, carry out the risk classification and prepare the catalogue of obligations and a usage policy for ChatGPT, Copilot and similar tools. You receive a documented and legally sound framework for the use of AI in your company, at a fixed price agreed in advance.

AI Check under the AI Act

Fixed price, agreed before work begins.

To the advisory modules →

Frequently asked questions

Yes. Anyone who uses AI systems in their own business is generally a deployer within the meaning of the regulation. For most everyday applications the obligations are manageable, but they start with the AI literacy of employees and extend to transparency obligations, for example when using a chatbot.

This article provides a general overview for information purposes only and does not replace advice on individual cases. As of: August 2026.

Are you using AI in a legally compliant way?

Briefly tell us which AI tools you use. We will get back to you within one business day with an initial assessment.