Workflows18 May 202613 min read
EU AI Act for SMEs: New Deadlines After the Omnibus
The AI Act high-risk rules apply only from 2 December 2027. What has applied since 2 August 2026 and what SMEs need to arrange now.
Co-founder

Introduction
The EU AI Act already applies in part: prohibited AI applications have been banned since February 2025, and since 2 August 2026 you must tell users when they are talking to AI or viewing AI-generated content. The heavy rules for high-risk AI, such as CV screening or credit scoring, have been postponed to 2 December 2027 by the Digital Omnibus (Regulation (EU) 2026/1744). For most SME use cases that means: arrange transparency now, plan high-risk documentation.
Last updated: 14 October 2026. What changed: the high-risk deadlines have shifted because of the Digital Omnibus (Annex III to 2 December 2027, Annex I to 2 August 2028). The table, checklist and frequently asked questions have been adjusted accordingly; the first version appeared on 18 May 2026.
What changes and when?
| What | Original date | New date | Status |
|---|---|---|---|
| Prohibited AI practices (Art. 5) | 2 Feb 2025 | unchanged | in force |
| Obligations for providers of general-purpose AI models | 2 Aug 2025 | unchanged | in force |
| Transparency duty for chatbots and AI content (Art. 50) | 2 Aug 2026 | unchanged | in force |
| High-risk AI under Annex III (incl. recruitment, credit, education) | 2 Aug 2026 | 2 Dec 2027 | postponed |
| High-risk AI in products under Annex I (e.g. machinery, medical devices) | 2 Aug 2027 | 2 Aug 2028 | postponed |
Source: Regulation (EU) 2026/1744, Official Journal of 24 July 2026, in force 27 July 2026. For providers of systems that generate synthetic audio, images, video or text and were already on the market before 2 August 2026, a transition period until 2 December 2026 applies to the machine-readable marking (Art. 50(2)).
Not sure which category your AI feature falls into? We will walk you through it.
What the AI Act actually regulates
The AI Act does not regulate technology. It regulates use cases. The same model can fall under minimal risk in one context and under high risk in another.
Summary of Article 6 and Annex III, EU AI Act
The AI Act is a European regulation that classifies AI systems by risk and attaches different obligations to each tier. Four categories: unacceptable risk (banned), high risk (heavily regulated), limited risk (transparency obligations) and minimal risk (no obligations). The Act looks at the purpose and impact of the system, not at the underlying technology. It does have to be an AI system, though: according to recital 12 of the regulation, systems that only execute rules defined in advance by people fall outside the definition. A machine learning model deciding on mortgage applications falls into the same category as a sophisticated neural network doing the same thing.
For SMEs that distinction is crucial. A chatbot answering FAQ questions falls under limited risk: you must make clear to the user that they are talking to an AI. A recommendation engine in your webshop typically falls under minimal risk: in principle no obligations. An algorithm screening job applicants or scoring customer credit falls under high risk: that is where the heaviest obligations begin.
When the Act really affects your SME
The Act defines two main roles: provider and deployer. A provider develops an AI system or has it developed and places it on the market under their own name or brand. A deployer uses an existing system within their own operations. Most Dutch SMEs are deployers. They use Microsoft Copilot, ChatGPT Enterprise, Salesforce Einstein or a Dutch SaaS with AI features. For deployers of low-risk systems, the Act is mainly a transparency and governance exercise.
It changes the moment you commission custom software in which AI plays a role. Then your builder potentially becomes a provider, and you potentially become a deployer of a high-risk system. It also changes if you substantially modify an existing tool or deploy it for a new purpose. A company that starts using a chatbot built for customer service to screen job applicants changes role under Article 25. You then become a provider yourself, with all the associated obligations. This is exactly the kind of transition SMEs slip into without a legal review.
For MG Software clients: if we build an AI feature in which we make or materially influence decisions for the end user, we design the system from the first sprint with provider obligations in mind. That prevents a lot of retrofitting of documentation later.
Annex III: when your system is high-risk
Annex III of the Act lists eight categories where AI systems automatically count as high-risk once they make or materially influence decisions. The most relevant for SMEs are: employment and HR, access to essential services such as credit and life or health insurance, biometric identification and, in some cases, education. Logistics optimization, inventory management, marketing segmentation and internal productivity tools usually fall outside Annex III.
A few concrete examples from our client base. A SaaS scoring CVs for recruiters: high-risk under Annex III section 4 (employment). A platform that uses machine learning to decide which customer qualifies for instalment payment: high-risk under Annex III section 5 (essential services, credit scoring). A dashboard that uses AI for sales forecasting for the buyer: not high-risk, because it influences no decision about a person. A customer service chatbot: limited risk, transparency obligation only.
The dangerous grey area sits in tools that hover at the edge. A personalised quote for life or health insurance computed by AI can fall under Annex III. A dynamic price calculator for a webshop does not. The difference is whether the outcome materially affects an individual in a domain the Act protects. When in doubt, document the reasoning and have it reviewed by a lawyer.
The documentation requirements for high-risk AI (from 2 December 2027)
For each high-risk system the Act prescribes seven concrete obligations, set out in Articles 9 through 15. For Annex III systems they apply from 2 December 2027, for systems in products under Annex I from 2 August 2028. They are not optional and they are interconnected. Building them in during construction is a workable amount of effort. Bolting them on after the fact to a system that is already running often takes weeks.
One: a risk management system that runs across the entire lifecycle (Article 9). The system must identify known and reasonably foreseeable risks and reconsider them periodically. Two: data governance (Article 10). You must be able to demonstrate where training, validation and test data come from, whether bias exists, how representativeness is ensured and how personal data has been handled. Three: technical documentation (Article 11). Annex IV provides a detailed table of contents: system architecture, design choices, hardware requirements, monitoring and evaluation metrics.
Four: automatic logging (Article 12). The system must record operational behaviour in a way that is traceable for regulators. Five: transparency to deployers (Article 13). The user manual must make clear what the system can and cannot do, what the known limitations are and how to use it correctly. Six: human oversight (Article 14). The architecture must make human intervention practically possible, including a kill switch. Seven: accuracy, robustness and cybersecurity (Article 15). The system must perform measurably, be resilient against errors and protected from attempts to manipulate its outputs.
For our clients we now deliver these seven blocks as a standard pack. Risk register in Markdown inside the repo, datasheets per dataset, a Notion or Confluence template for Annex IV, structured audit logs in a separate database, a user manual generated from code and docs, a human-in-the-loop interface where required, and automated tests for accuracy drift. Not sexy, but sellable to enterprise buyers and defensible in front of regulators.
Five typical SME scenarios and what to do
Scenario one: you use ChatGPT Enterprise for internal productivity. Limited risk. Arrange transparency for employees and set up an internal AI policy. No high-risk obligations. Time investment: limited.
Scenario two: you commission an AI chatbot for customer service. Limited risk, provided it does not decide on essential services. Show on the first interaction that it is AI. Log conversations for quality. Define an escalation path to a human. Time investment: moderate, mostly during design.
Scenario three: you build a SaaS where an AI component screens job applicants. High-risk. All seven obligations apply. For an Annex III system like this, the conformity assessment is done through internal control; a notified body is not needed. You then draw up an EU declaration of conformity and affix the CE marking. Register in the EU database for high-risk systems before putting the system into service. Time investment: large, ideally built in from the start.
Scenario four: you sell life or health insurance and use machine learning for per-customer premium calculation. High-risk. Document historical training data, monitor sensitive attributes, make premium decisions explainable to the customer. Work with the actuarial team on data quality. Time investment: substantial, particularly data governance.
Scenario five: you commission a logistics dashboard with AI for route optimization. Minimal risk, provided it does not evaluate individual workers. No obligations under the AI Act. Still wise to arrange explainability and monitoring, because client trust and auditability remain important.
Fines and what enforcement looks like in practice
The Act has a three-tier fine structure, set out in Article 99 of Regulation (EU) 2024/1689. Up to 35 million euros or 7 percent of global annual turnover for prohibited practices such as social scoring or manipulative techniques. Up to 15 million euros or 3 percent for failing to comply with high-risk obligations. Up to 7.5 million euros or 1 percent for providing incorrect information to regulators. For SMEs, including start-ups, the lower of the two amounts applies (Article 99(6)), but the concept is clear: serious violations are treated seriously.
More important than the theoretical maxima is what enforcement will look like in practice. That is not yet fully clear, because national supervision is still being set up. We expect complaints to become an important trigger for investigations, for example from a rejected job applicant or an unhappy customer. On top of that, in our experience enterprise buyers increasingly ask for proof of AI Act compliance. A software vendor without credible documentation can therefore lose tenders well before a fine comes into view.
For the Netherlands, the Dutch cabinet proposes a system in which several existing regulators cooperate and each supervise AI within their own domain. The Autoriteit Persoonsgegevens and the Rijksinspectie Digitale Infrastructuur (RDI) get a coordinating role. This concerns a bill that the cabinet put out for internet consultation on 20 April 2026. In practice, requests for documentation and proactive audits can therefore arrive via different routes.
Checklist: what to arrange now and what before December 2027
Work through the checklist per system, not once at organisation level. The first list applies to all AI you use or build, the second only to high-risk systems.
Now (already applies)
- Make an AI inventory of all tools and in-house features.
- Make chatbots and AI content recognisable to users.
- Check that no application falls under the prohibited practices.
- Give employees who work with AI basic training.
Before 2 December 2027 (high-risk only)
- Determine whether you are a provider or a deployer.
- Set up risk management and technical documentation.
- Build in logging and human oversight.
- Prepare registration and conformity assessment.
How we build AI Act compliant systems
At MG Software we start every AI project with a risk classification during the discovery phase. It is a short, well-defined step and determines whether we design for light compliance (transparency, logging) or heavy compliance (all seven obligations). Architecture, model selection and the data layer are tuned to that outcome. We do not build high-risk systems on a black-box API without logging capabilities. We choose models where we can record the inputs and outputs.
Documentation is not something we write at the end of a project. Annex IV fields are filled in during sprints, at the same moment the feature is built. The risk analysis sits in the same repo as the code. Datasheets sit next to the migrations. The user guide sits in the docs folder. At delivery, clients have both a working system and a compliance pack that survives an audit.
Do you have an existing AI system without clear documentation? Then we run a compliance audit with a concrete remediation plan and a timeline towards 2 December 2027. We map what is missing, estimate the effort and often implement the fixes. For a quick indication of scope and cost, use our project calculator or book a conversation directly.
Want to know what it costs first? Read the cost of an AI feature with figures from our projects. For systems that perform tasks independently, read more about agentic AI in building an AI agent for your business processes. And if you supply software to organisations that fall under the new security act, also read our explanation of the Cybersecurity Act for software suppliers.
Conclusion
The EU AI Act is not a reason to remove AI from your business. It is a reason to handle AI more maturely. For most SME use cases the burden is limited and the workload reasonable. For high-risk systems it is serious work, but work that pays back in more reliable systems and in credibility towards customers and partners.
The transparency duty has applied since 2 August 2026. The heavy high-risk obligations apply to Annex III systems from 2 December 2027. Those who start now have time to do it right. Compliance is not a cost if you build it in early. It is then simply part of good engineering. Want a system built that performs tasks independently? Then read how to approach building an AI agent for business processes, or get in touch for an AI Act readiness scan.
Sources (consulted 13 October 2026)
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), EUR-Lex, Official Journal of 24 July 2026.
- Regulation (EU) 2024/1689 (AI Act), EUR-Lex, including recital 12, Article 25, Article 50, Article 99 (fines) and Annex III.
- Rijksoverheid, news item of 20 April 2026: supervision of the AI Regulation in the Netherlands.
Frequently asked questions
Has the AI Act been postponed?
Partly. Only the high-risk rules have moved: Annex III to 2 December 2027, Annex I to 2 August 2028. Prohibited practices, GPAI rules and the transparency duty already apply.
Does my chatbot already have to comply?
Yes. Since 2 August 2026 it must be clear that users are communicating with an AI system.
When does my software count as high-risk?
When it falls into an Annex III category, such as recruitment and selection or credit scoring. Internal productivity, inventory or marketing tools usually do not.

Co-founder
Related posts
WorkflowsDutch Cybersecurity Act (NIS2): What Changes for SoftwareRead article
WorkflowsDigitalisation Subsidies 2026: How to Partly Fund Custom SoftwareRead article
WorkflowsFunctional design template for custom softwareRead article
WorkflowsThe Dutch Tax Authority Is Enforcing False Self-Employment Rules: What It Means for Software HiringRead article