Something shifted in the EU policy landscape in late 2025, and the organisations paying attention are already moving.
In October 2025, the European Commission published the Apply AI Strategy — the EU's overarching sectoral plan for AI adoption across industry and the public sector. A few months earlier, in April 2025, the Commission's AI Continent Action Plan announced the creation of the EU AI Skills Academy: a centralised hub that will aggregate AI and generative AI training from every major EU-funded programme — European Digital Innovation Hubs (EDIHs), AI Factories, the European Institute of Innovation and Technology (EIT), and more.
Both of these initiatives sit directly on top of a legal obligation that has been in force since 2 February 2025: Article 4 of the EU AI Act, which requires every organisation that deploys AI systems to ensure its staff maintain a sufficient level of AI literacy.
For the 26 million SMEs operating across the European Union, this combination of policy signal and legal mandate creates a clear and immediate challenge — and, for the businesses that move first, a significant competitive advantage.
This article explains what changed, what the law actually requires, where the skills gap stands today, and how your organisation can close it without hiring a data science team.
Table of Contents
- What Just Changed: The Policy Landscape in 2026
- What Article 4 Actually Requires — In Plain Language
- The Skills Gap Reality: 74% of EU SMEs Can't Find AI-Ready Staff
- The Infrastructure Signal: How the EU Is Building the Training Ecosystem
- The Practical Path: What Non-Technical Teams Can Do Right Now
- Getting Your Team Started with AI Agent Camp
- Frequently Asked Questions
1. What Just Changed: The Policy Landscape in 2026
The EU's approach to AI skills has evolved from aspiration to infrastructure in the space of twelve months. Three interlocking developments define where things stand today.
The Apply AI Strategy (October 2025)
Published in October 2025 by the European Commission, the Apply AI Strategy is the EU's overarching plan for AI adoption across ten key industrial sectors — from healthcare and manufacturing to agri-food and cultural industries — plus the public sector.
At its core, the strategy promotes what the Commission calls an "AI first policy": when organisations make strategic or operational decisions, AI should be assessed as a potential solution, with careful consideration of both benefits and risks. This isn't just aspirational language; the strategy contains concrete measures tied to funding programmes and governance mechanisms, including the newly established Apply AI Alliance — the coordination forum that brings together AI providers, industry leaders, academia, and the public sector.
For workforce development specifically, the Apply AI Strategy introduces three workforce priorities:
- AI literate workers for all sectors — ensuring that every worker in every sector has access to AI literacy training tailored to their job profile
- Sectoral AI experts for digitally intense sectors — creating multidisciplinary profiles who can develop AI applications for specific industries
- Active monitoring of AI's labour market impact — ongoing tracking of displacement risks, skill transition needs, and demographic disparities
The first priority directly affects your organisation today.
The EU AI Skills Academy (April 2025)
Launched as part of the AI Continent Action Plan in April 2025, the EU AI Skills Academy is designed to be the single point of access for AI and generative AI training across all EU-funded initiatives.
Rather than building a new institution from scratch, the Academy works as an aggregator: it will pull together training programmes from EDIHs, AI Factories, the EIT's Knowledge and Innovation Communities, and the Interoperable Europe Academy into one place. The Academy will also develop a pilot generative AI-focused degree, an AI apprenticeship programme, and scholarship and returnship schemes specifically aimed at attracting more women into the AI field.
For HR managers and L&D professionals at European SMEs, the practical implication is this: EU-subsidised training pathways are now being organised and made accessible in a way they haven't been before. The infrastructure is being built. The organisations that learn how to navigate it and supplement it with faster, more practical training will be ahead.
Article 4 of the EU AI Act (In Force Since 2 February 2025)
Article 4 didn't arrive with the Apply AI Strategy — it has been in force since 2 February 2025. But the launch of the Skills Academy and Apply AI Strategy marks the moment when the EU's enforcement and support machinery began to catch up with the legal requirement.
The practical reality: if your organisation deploys AI systems — and in 2026, almost every organisation does — Article 4 applies to you. We cover exactly what it requires in the next section.
2. What Article 4 Actually Requires — In Plain Language
Article 4 of the EU AI Act is often summarised as an "AI literacy" requirement, which is accurate but undersells its operational implications. Here is what it actually says:
"Providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf."
Let's unpack the key terms.
"Providers and deployers" — A provider builds or places an AI system on the market. A deployer uses an AI system in a professional context. If your company uses an AI tool for HR screening, customer service, financial analysis, or any other business function, you are a deployer. Most European SMEs fall into the deployer category.
"Shall take measures" — This is not aspirational. The requirement is to actively take steps, not just acknowledge the issue. National market surveillance authorities — the agencies responsible for enforcing the AI Act in each EU member state — are now in the remit for Article 4 supervision.
"Sufficient level of AI literacy" — The AI Act defines AI literacy as the skills, knowledge, and understanding that enables providers and deployers, as well as affected persons, to make an informed deployment of AI systems. Critically, the law specifies that providers and deployers must consider the technical knowledge, experience, education, and training of their staff, as well as the context in which AI is used and the persons who may be affected by AI outputs.
"Staff and other persons dealing with AI on their behalf" — This extends beyond your direct employees to contractors, partners, and service providers who interact with your AI systems.
What Compliance Looks Like in Practice
The EU AI Office — the body responsible for supporting implementation — published a Questions & Answers on AI literacy that provides guidance on what "sufficient level" means in context. It also launched a Repository of AI Literacy Practices containing more than 40 real examples submitted by companies and public sector organisations.
Importantly, the AI Office states that replicating practices from the repository does not automatically grant presumption of compliance. Each organisation must assess its own context — which AI systems it deploys, who uses them, what the stakes are, and what knowledge gaps exist.
For most SMEs deploying commercially available AI tools (AI-powered CRM, automated document processing, AI-assisted HR screening), compliance with Article 4 means:
- Inventorying AI tools in use across the organisation
- Identifying which staff interact with those tools in their professional capacity
- Assessing current AI literacy levels against the context of each tool's use
- Implementing training proportionate to the role and risk level
- Documenting the training provided
The good news: Article 4 is explicitly proportionate. A small company deploying a low-risk AI customer service chatbot does not need the same literacy programme as a financial institution using AI for credit scoring. But "we haven't done anything yet" is not a proportionate response to the legal obligation — it's a compliance gap.
3. The Skills Gap Reality: 74% of EU SMEs Can't Find AI-Ready Staff
Policy urgency is one thing. Market reality is another — and the data for European SMEs is stark.
According to the Fortune/Google "AI Works for Europe" report (March 2026), 74% of EU SME employers report they cannot find workers with adequate AI skills to fill their current needs. This is not a future problem. It is a present-day operational constraint.
The same research highlights a sector-specific shock: demand for AI skills in Accounting and Finance has tripled since 2023. Finance professionals — the people managing company cash flows, invoicing, payroll, and compliance reporting — are among the fastest-growing targets for AI integration, and the training supply has not kept pace with the demand signal.
Several interconnected factors drive this gap:
The speed of tooling outpaces training. AI tools are being embedded into business software faster than organisations can build internal capability. Microsoft Copilot is integrated into Office 365. Salesforce Einstein is built into CRM workflows. HubSpot AI is live in marketing platforms. The tools are there; the understanding of how to use them effectively and responsibly is not.
Technical training dominates available supply. Most AI training programmes available in the market are designed for developers, data scientists, or engineers. HR managers who need to understand AI-assisted resume screening, or finance teams navigating AI-generated forecasts, or operations directors deploying process automation agents — these professionals have had limited options for practical, role-appropriate training.
Compliance creates new demand. Before Article 4, AI literacy training was a competitive nice-to-have. Now it is a legal obligation. This shift in framing changes the calculus for HR budgets: training is no longer a discretionary investment; it is a compliance cost — and doing it well creates organisational capability that also drives performance.
The 74% figure from the AI Works for Europe report is a snapshot of today's unmet need. With the Apply AI Strategy pushing AI adoption across ten sectors, and the Skills Academy beginning to aggregate training resources, the gap will get more visible before it gets smaller.
4. The Infrastructure Signal: How the EU Is Building the Training Ecosystem
Understanding the EU's AI skills infrastructure matters for two reasons: it tells you what publicly funded resources will be available, and it signals how quickly the regulatory and competitive environment will intensify.
European Digital Innovation Hubs (EDIHs): Transformed Into Experience Centres for AI
The 230+ EDIHs operating across the EU have been redefined by the Apply AI Strategy. Under the new framework, EDIHs have been transformed into Experience Centres for AI — access points to the broader EU AI innovation ecosystem, including AI Factories, Testing and Experimentation Facilities, and AI regulatory sandboxes.
For SMEs, EDIHs remain the most accessible entry point for publicly supported AI assistance. Many offer subsidised "test before invest" services — allowing SMEs to pilot AI tools with technical support before committing resources. The transformation to Experience Centres adds a skills and training dimension that was previously inconsistent across the network.
Finding your nearest EDIH: the European Digital Innovation Hubs catalogue is available at european-digital-innovation-hubs.ec.europa.eu.
AI Factories: Compute Access for Experimentation
The EU's network of AI Factories provides access to high-performance computing infrastructure — the kind needed to train and fine-tune AI models. For most SMEs, AI Factories are less immediately relevant than EDIHs, but they are an important part of the ecosystem for organisations building custom AI capabilities.
Under the Apply AI Strategy, AI Factories are also expected to contribute training content to the EU AI Skills Academy, creating a pipeline from advanced research capability to practical workforce education.
The EU AI Skills Academy: When It Arrives, It Will Aggregate — Not Replace
The EU AI Skills Academy is currently in development phase. The AI Continent Action Plan announced it in April 2025, and the first funding calls opened under the DIGITAL programme (reference: DIGITAL-2025-SKILLS-08-GENAI-ACADEMY-STEP). Full rollout of aggregated courses across all EU-funded sectoral initiatives will take time — likely 2026–2027 for the broadest content library.
This means the skills gap your organisation faces today will not be closed by the Skills Academy in the near term. The Academy will be a valuable long-term resource, but businesses that wait for it to be fully operational before beginning AI literacy efforts will arrive late to a compliance obligation that is already active.
The AI Office's AI Literacy Repository: A Free Resource Available Now
One resource that is already live: the EU AI Office's Repository of AI Literacy Practices. This repository contains more than 40 documented examples of how companies and public sector organisations have approached AI literacy — including self-assessment frameworks, role-based training modules, and awareness campaigns.
The repository is a useful benchmark. It does not provide ready-to-deploy training, but it offers concrete examples of what compliant, well-structured AI literacy programmes look like in practice.
What This Means for SMEs That Wait
The EU's AI skills infrastructure is being built — but it will take years to reach its full scope. Meanwhile:
- Article 4 is already in force and being enforced by national authorities
- The 74% skills gap in the SME market is a present operational constraint
- Competitors that build internal AI capability now will widen their advantage as the Apply AI Strategy accelerates adoption
The organisations that use the EU's infrastructure as a supplement — while pursuing practical, immediately deployable training today — will be better positioned than those waiting for a single EU programme to solve the problem.
5. The Practical Path: What Non-Technical Teams Can Do Right Now
For HR managers, L&D professionals, and operations directors at European SMEs, here is a structured approach to closing the AI literacy gap in a way that is proportionate to your organisation's size and risk profile.
Step 1: Map Your AI Exposure
Before designing a training programme, you need to know what you're training for. Create a simple inventory:
| AI Tool in Use | Department | Users | Role in Decision-Making |
|---|---|---|---|
| AI-powered CRM | Sales | Sales reps, managers | Lead scoring, outreach drafting |
| Document processing | Finance | AP team | Invoice review, payment approvals |
| HR screening tool | People | Recruiters | CV shortlisting |
| Analytics platform | Marketing | Marketing team | Campaign optimisation |
This inventory — which you may already have in part from GDPR data mapping exercises — tells you which staff fall under Article 4's scope and what the risk context is for each tool.
Step 2: Assess Current Literacy Levels
Literacy assessment doesn't need to be complicated. A short structured questionnaire covering:
- Understanding of what AI outputs are and aren't (probabilistic vs. deterministic)
- Awareness of bias and errors in AI systems
- Knowledge of when to override, escalate, or verify AI recommendations
- Understanding of data privacy implications of AI tool use
The EU AI Office's Q&A guidance provides a useful framework for what adequate literacy looks like in different deployment contexts.
Step 3: Design Role-Appropriate Training
Article 4 is explicit that training should be proportionate to context. A finance professional approving AI-recommended payment decisions needs deeper training than a marketing manager using AI to draft subject line variations.
Practical training should cover:
- What the AI tool does — not the underlying architecture, but the decision logic and output format
- Where it can be wrong — known limitations and failure modes
- How to verify and override — practical workflows for human-in-the-loop oversight
- What data it uses — GDPR implications of inputs and outputs
- When to escalate — clear criteria for routing decisions to a human
Step 4: Build Capacity for Agents Specifically
The AI landscape in 2026 is moving rapidly toward agentic AI — AI systems that don't just generate text or classify inputs, but take sequences of autonomous actions on behalf of users. AI agents draft and send emails, query databases, schedule meetings, process invoices, and update records without step-by-step human instruction.
The Apply AI Strategy's focus on 10 key sectors and the Skills Academy's emphasis on generative AI training both signal that agentic AI — not just chatbots — is where enterprise AI adoption is heading. Finance, HR, customer service, and operations are precisely the functions where agentic AI is landing fastest.
For non-technical teams, the most valuable AI literacy investment in 2026 is understanding:
- How to configure AI agents for specific business workflows
- How to set appropriate oversight checkpoints
- How to evaluate agent outputs and correct course when agents make errors
- How to document agent use in a way that satisfies Article 4 obligations
This is a practical skillset — not a theoretical one. It can be learned without a programming background. And it directly maps to the compliance obligation Article 4 creates.
Step 5: Document Everything
Compliance with Article 4 requires not just training but documentation. Maintain records of:
- Which staff received training, and when
- What training content was covered, relevant to which AI tools
- Assessment results (if formal assessment was conducted)
- Refresher schedules as AI tools are updated or added
This documentation serves two purposes: it demonstrates good-faith compliance to national market surveillance authorities if requested, and it creates an internal knowledge baseline that makes future training more efficient.
6. Getting Your Team Started with AI Agent Camp
The EU AI Skills Academy will eventually provide aggregated access to EU-funded training. EDIH Experience Centres will expand their skills services. These are long-term infrastructure investments.
But your Article 4 obligation exists today. And the 74% skills gap in the EU SME market is happening now, quarter by quarter, as AI tools become standard across every business function.
AI Agent Camp offers European SMEs a practical, immediately deployable solution:
- Self-paced curriculum designed specifically for non-technical professionals — HR managers, L&D leads, finance teams, operations directors
- Focus on AI agents — the most strategically relevant AI capability for business professionals in 2026, aligned with the Apply AI Strategy's workforce priorities
- No-code approach — participants learn to design, configure, and oversee AI agents without writing a single line of code
- Article 4-aligned content — the curriculum covers AI literacy fundamentals, bias and error awareness, human oversight design, and data privacy — the exact domains that constitute "sufficient AI literacy" under the AI Act
- $89/month — accessible for individual professionals and scalable for teams
Whether your team is starting from zero or building on existing digital skills, AI Agent Camp provides the practical foundation needed to meet Article 4's requirements and capture the productivity benefits that the Apply AI Strategy is designed to unlock.
→ Start Your Team's AI Agent Upskilling Today at AI Agent Camp
No technical background required. Self-paced. Cancel anytime. $89/mo.
7. Frequently Asked Questions
Q: Does Article 4 of the EU AI Act apply to SMEs, or only large enterprises?
Article 4 applies to all deployers of AI systems — including SMEs. There is no size threshold. However, the proportionality principle means that the level of training required scales with the risk profile of the AI tools deployed and the nature of their use. An SME using a low-risk AI scheduling tool has a lighter obligation than a financial institution using AI for credit decisions. But the obligation to take some measure exists regardless of company size.
Q: When does the EU AI Skills Academy open for enrolment?
The EU AI Skills Academy is currently being developed under the DIGITAL-2025-SKILLS-08 funding call. The first cohorts are expected as the programme reaches its operational phase in 2026–2027. The Skills Academy will aggregate content from EDIHs, AI Factories, EIT and other EU-funded initiatives. For training needs that exist today, SMEs should not wait for the Academy's full launch.
Q: How do EDIHs relate to AI literacy training under the Apply AI Strategy?
Under the Apply AI Strategy, EDIHs have been transformed into "Experience Centres for AI." They are expected to expand their skills and training services related to AI and will contribute training content to the EU AI Skills Academy. Many EDIHs already offer subsidised training and "test before invest" services for SMEs. Contact your nearest EDIH to understand what is available in your region: european-digital-innovation-hubs.ec.europa.eu.
Q: What is the enforcement mechanism for Article 4?
Supervision and enforcement of Article 4 falls to national market surveillance authorities in each EU member state — the same agencies that enforce other market regulations. The AI Office supports implementation by working with organisations and member states through the AI Board, but national authorities have the enforcement mandate. This means enforcement will vary by country and context, but the legal obligation is uniform across the EU.
Q: Is AI literacy training the same as AI technical training?
No. Article 4 requires AI literacy — which the AI Act defines as the skills, knowledge, and understanding that enable informed deployment and oversight of AI systems. This is explicitly distinguished from technical expertise. A finance professional needs to understand what an AI-assisted invoice tool does, where it can fail, and when to escalate — not how the neural network was trained. Practical, role-focused AI literacy training is sufficient for Article 4 compliance for most SME roles. Technical depth is required only for staff directly involved in developing or configuring AI systems.
Q: Can my team take AI Agent Camp training as part of our Article 4 compliance programme?
Yes. AI Agent Camp's curriculum covers the core domains of AI literacy defined in the AI Act and the AI Office's Q&A guidance: understanding how AI systems work, awareness of limitations and potential errors, human oversight practices, and data privacy considerations. Combined with documentation of training participation and relevance to specific AI tools your organisation deploys, this constitutes a reasonable compliance measure under Article 4's proportionality principle. We recommend maintaining training records as part of your broader AI governance documentation.
The Bottom Line
The EU AI Skills Academy and Apply AI Strategy are not distant policy announcements. They are the visible tip of a regulatory and infrastructure shift that is already active, already funded, and already shaping the competitive landscape for European SMEs.
Article 4 of the EU AI Act has been in force since February 2025. The 74% skills gap documented in the "AI Works for Europe" report is a present operational reality. The Apply AI Strategy's push across ten sectors will accelerate AI adoption — and with it, the gap between organisations whose workforce is ready and those that aren't.
The organisations that act now — mapping their AI exposure, building role-appropriate literacy programmes, and developing internal capacity for the agentic AI tools that are already entering their business software — will meet their compliance obligations and gain a durable advantage over competitors that are still evaluating whether AI training is worth the investment.
The EU is building the infrastructure. The policy signal is clear. The practical path forward starts today.
→ Start Your Team's AI Agent Upskilling at AI Agent Camp — $89/mo
Sources and Further Reading
- EU AI Act, Article 4 — AI literacy obligation for providers and deployers. Official text via EUR-Lex
- EU Apply AI Strategy (October 2025) — digital-strategy.ec.europa.eu/en/policies/apply-ai
- AI Continent Action Plan (April 2025) — digital-strategy.ec.europa.eu/en/library/ai-continent-action-plan
- AI Talent, Skills and Literacy — EU policy hub — digital-strategy.ec.europa.eu/en/policies/ai-talent-skills-and-literacy
- AI Literacy — Questions & Answers — digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers
- Repository of AI Literacy Practices — digital-strategy.ec.europa.eu/en/policies/repository-ai-literacy-practices
- Fortune/Google "AI Works for Europe" Report (March 2026) — 74% EU SME employers; Finance AI demand tripled
- European Digital Innovation Hubs Catalogue — european-digital-innovation-hubs.ec.europa.eu
Last updated: April 2026. Policy information sourced from official EU Commission publications. SME statistics from Fortune/Google "AI Works for Europe" report (March 2026).
Ready to put AI agents to work?
Turn what you just read into real workflows. AI Agent Camp helps non-technical professionals go from using to building — hands-on.
Last reviewed: 2026-05-30