AI literacy is now mandatory: how Dutch SMEs can use the SLIM subsidy to comply

Key Takeaways: Starting August 2, 2026, the European AI Act legally requires businesses that deploy AI systems to ensure their staff has sufficient AI literacy. The Dutch SLIM subsidy offers SMEs up to 25,000 euros (60% of costs) to fund exactly this type of training. Round 1 is open from April 7 through May 4, 2026, the window to act is now.

The AI Act deadline is approaching: what it means for Dutch businesses

On August 2, 2026, Article 4 of the European AI Act takes effect. This provision mandates that all providers and deployers of AI systems must ensure their personnel possess sufficient AI literacy. This is not a soft recommendation. It is a binding legal obligation with meaningful consequences for non-compliance.

For many Dutch businesses, this deadline arrives at a time when AI adoption is still very much a work in progress. Only 22.7% of Dutch companies with more than ten employees currently use AI. Among non-adopters, 73% cite a lack of knowledge and experience as the primary barrier. The gap between the legal requirement and current practice is substantial, but entirely bridgeable. The key lies in a combination that remarkably few businesses in the Netherlands have recognized: linking the AI Act obligation to the SLIM subsidy, which is specifically designed to fund workforce training for SMEs.

This article shows how you can address two challenges at once. You comply with the upcoming legislation while simultaneously strengthening your team's AI capabilities, with significant financial support from the Dutch government.

What is AI literacy and why is it now legally required?

The definition under the AI Act

Article 4 of the AI Act defines AI literacy as the ability of personnel to make informed decisions about AI systems. In practice, this encompasses three competencies: understanding what AI systems do and how they work at a level appropriate to one's role, recognizing and assessing the risks and limitations of AI, and evaluating the impact of AI output on decision-making and business processes.

The legislator acknowledges that not everyone needs the same depth of knowledge. A marketing professional reviewing AI-generated content requires different competencies than a developer integrating AI models into business software. The obligation stipulates that the level of literacy must be proportional to the role, the technical background of the individual, and the context in which the AI system is deployed.

Why Europe is taking this step

The motivation behind this obligation is anything but abstract. Europe observes that AI systems are being rapidly deployed in critical business processes, from recruitment and selection to credit assessment and medical diagnostics. Without adequate knowledge among users, there is a real risk that AI output is accepted uncritically, that bias goes undetected, and that companies inadvertently violate the law through discriminatory or unsafe AI applications.

Real-world examples abound. A recruitment system that inadvertently disadvantages women because it was trained on historical hiring decisions. A chatbot providing customers with incorrect legal information without staff recognizing the errors. A prediction model rejecting credit applications based on postal codes, resulting in indirect discrimination. In each of these cases, a properly trained employee could have flagged the problem before it caused harm.

The consequences of non-compliance are not trivial. While Article 4 falls in the lowest risk category and the fines are less severe than for high-risk AI systems, supervisory authorities can and will enforce compliance. The Dutch Data Protection Authority (Autoriteit Persoonsgegevens) has been designated as the market surveillance authority for the AI Act in the Netherlands and has indicated it will actively enforce the regulation. More importantly, companies that demonstrably handle AI carelessly face reputational and liability risks that far exceed any direct fines. The AI Act is also part of a broader European regulatory framework that will only become more stringent in the years ahead.

Who falls under this obligation?

The scope is broad. The obligation applies to any organization that provides or deploys AI systems, regardless of size or sector. Does your company use ChatGPT for customer service, AI-powered accounting software, or automated marketing tools? Then you fall under Article 4. This encompasses not just technology companies, but every SME owner who uses AI tools in daily operations. Given the rapid adoption of generative AI tools, this affects a fast-growing number of businesses.

The SLIM subsidy: up to 25,000 euros for AI training

What is SLIM exactly?

The Stimuleringsregeling Leren en ontwikkelen in het MKB, known as SLIM, is a subsidy program from the Dutch Ministry of Social Affairs and Employment. The program is specifically designed to support SMEs in investing in employee development. With a total budget of 71.9 million euros for 2026, it is one of the largest subsidy pools for SME training in the Netherlands.

SLIM reimburses up to 60% of costs incurred for establishing training activities, with a maximum of 25,000 euros per application for individual SMEs. Collaborative partnerships of multiple SMEs can apply for up to 500,000 euros. This makes it particularly attractive to set up an AI literacy program jointly with industry peers.

Who qualifies?

The SLIM subsidy is available to SME businesses with up to 250 employees. Agricultural businesses and collaborative partnerships in hospitality, recreation, and agriculture also qualify. Foundations and associations that do not primarily operate as a business are excluded.

There are three initiative types for which you can request funding. The first involves assessing your organization to develop a future-oriented training and development plan. The second involves developing or adapting a methodology to improve employee skills. The third involves offering practical training positions for vocational education. AI literacy training fits into the first and second types.

Why SLIM and the AI Act are the perfect combination

Here lies the opportunity that virtually no one in the Netherlands is currently leveraging. The SLIM subsidy is designed for precisely the type of training that the AI Act now legally mandates. You can use the subsidy to establish an AI training plan that simultaneously ensures your compliance with the AI Act and genuinely strengthens your team's AI capabilities. Instead of viewing the AI Act as a cost center, it becomes the catalyst for a subsidized investment in your organization.

How to apply for SLIM: practical steps for round 1

Timeline and deadlines

SLIM operates with application windows. For 2026, there are two rounds. The first round runs from April 7 through May 4, 2026. The second round runs from August 10 through September 7, 2026. Applications are processed on a first-come, first-served basis, so early submission increases your chances.

Step by step through the application process

Applications are submitted through the digital portal of Uitvoering van Beleid (formerly RVO). You need eHerkenning level 3 to log in. Arrange this well in advance, as the eHerkenning application process can take several business days.

In the application, you describe the initiative for which you are requesting funding. For an AI literacy program, you outline which employees will be trained, which AI competencies you aim to develop, how the program is structured, and what costs you expect to incur. Be specific and tie activities to measurable objectives.

Eligible costs include direct expenses for developing or purchasing training materials and programs, external trainers or coaches, and the hours employees spend on training activities (at a fixed hourly rate). Non-eligible costs include regular salary expenses and overhead that cannot be directly attributed to the project.

Tips for a strong application

Make the connection to the AI Act explicit in your application. The subsidy provider values applicants who demonstrate that their training responds to concrete external developments and legal obligations. Describe how the program contributes to the sustainable employability of your staff in a changing technological landscape. If possible, include a quote from a recognized training provider and substantiate your budget with specific rates and hour estimates.

A common mistake is formulating outcomes too vaguely. Do not write "employees will learn about AI." Instead, specify: "Upon completion, 15 employees will be able to assess AI-generated reports for factual accuracy, understand which company data may and may not be entered into AI tools, and independently evaluate whether AI output meets the organization's quality standards." The more concretely you describe the training outcomes, the stronger your application will score in the assessment process.

What AI literacy training looks like in practice

A layered approach by role

Effective AI literacy is not a one-off half-day workshop. It is a structured program that accounts for the different roles and responsibilities within your organization. One approach that works uses three tiers.

The foundational tier targets all employees and covers the basics: what AI is, how language models and image recognition work at a high level, which AI tools the company uses and why, and what the basic principles of responsible AI use are. This includes recognizing hallucinations in AI output, understanding why confidential data should not be entered into public AI tools, and critically evaluating AI-generated suggestions.

The advanced tier targets employees who work with AI tools on a daily basis. This covers effective prompting, evaluating AI output for quality and bias, integrating AI into existing workflows, and documenting AI usage in accordance with AI Act requirements. A marketing team, for example, learns to systematically review AI-generated content for factual accuracy and brand consistency, while a finance team learns to validate AI analyses against source data.

The strategic tier targets management and decision-makers. This encompasses the legal frameworks of the AI Act, developing internal AI policies, risk assessment when procuring new AI tools, and establishing governance structures for responsible AI use.

Examples across sectors

In professional services, where much of the work is knowledge-based, AI literacy concentrates on the use of generative AI for text, analysis, and presentations. An accounting firm trains staff in using AI for financial statement analysis, but also teaches them to recognize the boundaries: when is human expertise indispensable, and how do you document AI-assisted conclusions?

In manufacturing, the emphasis lies on AI in production and quality processes. Operators learn how computer vision systems detect defects, why the system sometimes produces false positives, and how to respond to edge cases. Production managers learn to interpret the data behind predictive maintenance and assess the reliability of predictions.

In retail and e-commerce, AI literacy revolves around personalization algorithms, demand forecasting, and chatbots. Buyers learn how AI-driven demand forecasting works and when to override the algorithm. Customer service representatives learn to collaborate effectively with AI chatbots and when to intervene.

In logistics, the focus is on route optimization, warehouse management, and supply chain predictions. Planners learn to interpret AI suggestions for route planning and understand why the system makes certain choices, enabling them to make informed deviations when the situation calls for it.

Realistically estimating the time investment

A complete AI literacy program for an organization of 20 to 50 employees typically involves 40 to 80 hours of training time spread over two to three months. The foundational tier requires two to four hours per employee. The advanced tier demands eight to sixteen hours per employee, including hands-on exercises with the AI tools used in the business. The strategic tier for management spans four to eight hours. Additionally, time is needed for program preparation, developing company-specific training materials, and a post-program evaluation to verify whether learning objectives have been achieved. All of this fits comfortably within the SLIM subsidy framework.

Combining SLIM with other subsidies for broader AI adoption

WBSO for AI development

The SLIM subsidy covers the training component of AI adoption, but many businesses want to go further than training alone. For companies that develop their own AI solutions or innovatively apply existing AI technology in their business processes, the WBSO (Research and Development Tax Credit) offers a complementary fiscal incentive. The WBSO reduces labor costs for R&D activities and is particularly suitable for companies training AI models, building AI integrations, or developing data analysis pipelines.

The combination works: use SLIM to equip your team with the necessary AI knowledge, then deploy WBSO to make developing AI applications with that trained team fiscally advantageous. This way, you build a solid foundation of knowledge and immediately apply that knowledge in innovative projects.

Subsidies for AI implementation projects

Beyond training and R&D, there are specific subsidy opportunities for AI implementation projects. Consider the MIT program (SME Innovation Stimulus for Top Sectors), which subsidizes feasibility studies and collaborative R&D projects, or regional innovation funds that support investments in new technology. By strategically combining multiple programs, you maximize the financial support available for your entire AI adoption journey.

An integrated subsidy strategy

The most effective approach is to view AI literacy not as a standalone compliance project, but as the first step in a broader AI adoption trajectory. Start with a SLIM-funded training program that gives your team the foundation. Use the acquired knowledge to identify concrete AI applications that add value to your business processes. Finance the development of those applications with WBSO and other innovation subsidies. This layered approach spreads costs and risks while building step by step toward a data-driven organization.

Conclusion: three actions for the next four weeks

The AI Act deadline of August 2, 2026, is no longer an abstract future scenario. It is a concrete legal obligation taking effect in less than four months. At the same time, the first SLIM application round is open until May 4, creating a unique window to turn the obligation into an opportunity.

This week, inventory which AI systems your company uses and which employees work with them. Think broader than the obvious tools: beyond ChatGPT and Copilot, AI features embedded in existing software also fall under the AI Act, from smart search functions in your CRM to automated categorization in your accounting software. Map the current knowledge level and identify the largest gaps relative to what the AI Act requires.

Then draft a concise training plan that addresses the three tiers of AI literacy, tailored to the roles in your organization. Involve both the employees who work with AI daily and the management that makes decisions about AI system deployment. Finally, submit a SLIM application before May 4 to fund this training plan.

The businesses that act now achieve two things at once: they meet the legal requirements and they build the competencies needed to make AI genuinely valuable. Waiting until the deadline is not only risky from a compliance perspective, it also means leaving available subsidy funding on the table. With 71.9 million euros in SLIM budget and a growing number of applicants, early action is also strategically wise. If you miss round 1, mark August 10 as the start date for round 2.

Need support developing an AI training plan or submitting a SLIM subsidy application? Stratalytic guides SMEs through subsidy processes for AI and data innovation, from drafting the application to selecting the right training partners.