AI Act, Article 4 AI literacy In-company or online By profession

AI course for companies: from AI literacy to concrete impact per team

AI Act Article 4: why your employer has had to provide AI literacy since February

Since 2 February 2025, the European AI Act (Article 4) has required every organisation that uses AI systems to ensure its staff have an appropriate level of AI literacy. That applies to everyone who works with AI, from marketers using ChatGPT to finance teams using forecasting models to HR teams screening candidates. Our AI course is not a one-off ChatGPT workshop but a structured programme that builds AI fluency for each profession while mapping the impact areas for implementation.

Request a quote View the modules
AI-fluency Prompt-skills Tool-stack

Who this programme is for

Our AI course is aimed at organisations that want to go beyond the occasional ChatGPT tip. We work with directors and HR managers who want to prepare their teams properly for a workplace where AI is a standard part of day-to-day processes, and who also want to identify where AI implementation will deliver the most value.

The trigger is almost always a combination of three things. First, the legal obligation: since February 2025, the AI Act has required providers and deployers of AI systems to ensure their staff have sufficient AI knowledge, proportionate to their role and the context in which AI is used. Second, business urgency: employees are already using public AI tools, often without any framework for what is and isn't allowed with customer or company data. And third, the strategic question: where are the real productivity gains, and how do you uncover them when the organisation itself doesn't yet know what AI can do?

We deliberately don't position ourselves as a training provider selling standalone courses. We are an AI implementation partner: for many clients, the course is the starting point of a longer engagement in which we go on to build RAG applications, copilots, agent workflows or governance frameworks. The course delivers the literacy and the impact report your management team needs to decide, on solid grounds, where AI investment will deliver the highest return.

Modular offering by profession

An AI course that is identical for the whole organisation misses the point. A marketer needs different prompt skills and different tools than a controller or a software developer. That's why we build a separate module for each profession: 4 to 8 hours per group, 30% theory and 70% hands-on work with tools and data from the participants' own workflow.

Sales

Sales AI and account intelligence

Automating lead research with deep-research tools, generating account plans based on public signals, summarising call transcripts and filling CRM fields. In this module, participants work on their own account portfolio and finish with concrete prompt templates that can be used straight away in HubSpot, Salesforce or Pipedrive.

Marketing

Content AI and campaign acceleration

Turning briefs into content building blocks, SEO research with AI tools, image generation within your brand style, and adapting social content for each channel. We also cover the risks: AI content that dilutes your brand voice, and how to prevent that with style guides and review loops.

Customer service

Agent AI for support and the inbox

AI agents that classify tickets, prepare replies and retrieve knowledge articles. We show how a service team moves from first-line to second-line work when routine questions are handled by an agent, and which escalation rules are needed to safeguard quality.

Finance

Forecasting and data analysis

Speeding up spreadsheet work, setting up scenario analyses, detecting anomalies in transaction data and substantiating cash flow forecasts. We train finance staff to interpret AI output correctly and to build controls, so that AI remains a support tool and doesn't become a black box the management team follows blindly.

HR

Sourcing, screening and internal mobility

Optimising job ads, speeding up LinkedIn sourcing queries, and preparing candidate screening without the bias pitfalls that the AI Act explicitly classifies as high risk. In HR modules we pay extra attention to transparency and to keeping a human in the loop for selection decisions.

Operations

Process AI and workflow automation

Speeding up repetitive admin with copilots, mapping processes with AI, and setting up integrations with Make or n8n. In this module, participants identify concrete processes in their own department where AI automation saves the most time, with an initial impact estimate for each process.

Legal

Contract AI and compliance

Contract analysis with AI, faster due diligence and support for legal research. We cover how to use AI safely with confidential documents, which tools comply with professional secrecy, and how legal professionals check AI output without losing quality.

Engineering

Copilots, RAG and MCP for developers

Working effectively with GitHub Copilot, Cursor or Claude Code, building RAG applications on internal documentation, and using the Model Context Protocol (MCP) to give AI agents access to internal tools. For engineering teams that want to go beyond code completion.

Management

Governance and AI strategy for management

For directors and management team members: translating AI Act obligations into policy, setting up an AI register, risk classification per use case, budget allocation and assessing AI investment proposals. Not tool training, but the lens you need to frame AI decisions.

AI Act Article 4 in plain language

The AI Act has been in force since 1 August 2024. Article 4, the AI literacy obligation, is one of the first obligations to take effect, on 2 February 2025. This obligation affects virtually every organisation in the Netherlands.

The legislation states that providers and deployers of AI systems must "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". The level must match their technical knowledge, experience, education and training, the context in which the AI is used, and the persons or groups on whom the AI systems are used.

In practice, this means a marketing team using ChatGPT for copy needs a different level of literacy than an HR team using AI in recruitment procedures (a high-risk application under the AI Act) or an engineering team building AI agents with access to production systems. The law doesn't prescribe a specific curriculum, but it does expect you to be able to demonstrate that you acted proportionately.

Non-compliance falls under the broader AI Act enforcement framework. More important than the direct threat of fines is that Article 4 serves in practice as a benchmark: if something goes wrong with an AI application, such as an incorrect automated decision, a data leak through a prompt or a liability claim, the regulator or the court will look at whether you reasonably ensured your people knew what they were working with. A structured training programme with certification per employee is the evidence for that.

What we record for you: a certificate for each course participant with module content, date and competencies achieved. At organisation level, an AI literacy matrix showing which role has reached which level. This document fits into your AI register and can be used directly as evidence in audits, certification or incident investigations.

Sector-specific modules

In addition to role-based modules, we offer sector-specific deep dives. AI in healthcare calls for a different lens than AI in retail or financial services. Compliance requirements, data sensitivity, customer expectations and use cases differ per sector, and we tailor the content accordingly.

Healthcare and welfare

AI for administrative workload (record-keeping, claims), triage support and internal knowledge access for healthcare professionals. We cover the specific sensitivity of patient data, the NEN 7510 requirements and the difference between supporting AI and high-risk applications under the AI Act, such as diagnostic systems.

Finance and insurance

AI in fraud detection, customer contact, underwriting and compliance monitoring. Financial institutions subject to DORA and the AI Act face additional requirements around model validation, audit trails and explainability. Our module covers how you can deploy AI tools without putting your supervisory status at risk.

Retail and e-commerce

Product descriptions, customer segmentation, dynamic pricing, inventory forecasting and conversational commerce. We show how retailers use AI tools in marketing and logistics without compromising brand voice or legal consumer protection.

Legal and notarial

Specifically for law firms, notary offices and in-house legal departments: contract analysis, due diligence and legal research with retrieval-augmented generation. We cover the requirements arising from professional secrecy and the NOvA guidelines on AI use.

SMEs and business services

For smaller organisations without their own IT department: getting the basics right with secure tool choices, a workable AI policy and an initial impact scan. Practical and without unnecessarily heavy governance, yet fully AI Act compliant.

Industry and logistics

Predictive maintenance, quality control with computer vision, planning optimisation and logistics forecasting. We align AI applications with the operational reality of the factory floor, warehouse and transport planning.

In-company or online: the format that suits you

We train groups of 5 to 50+ employees, in batches that work for your organisation. Two formats, the same content and the same hands-on approach.

In-company

On-site at your organisation

A trainer comes to your office and works with your team on real cases from your own workflow. Suitable for organisations that want to get their teams up and running in a single day or a series of half-days. Ideal for department training where participants work on the same processes together, such as sales, marketing or customer service. We recommend in-company training when there is already a shared context and hands-on collaboration boosts learning.

In practice: 4 to 8 hours per module, in batches of 8 to 20 participants per group. We split larger companies into consecutive batches so the format stays intensive.

Digital

Live via Teams or Zoom

Live sessions on your own video platform: Microsoft Teams, Zoom or Google Meet. Suitable for organisations with multiple locations, hybrid teams or dispersed workplaces. Participants work in their own environment with their own tools, which speeds up the transfer to day-to-day work. Breakout rooms and live coaching preserve the hands-on dynamic.

In practice: sessions of 2 to 3 hours, often spread over several days or weeks. Recordings are available for anyone who cannot attend live, plus asynchronous Q&A in the intervening period.

We scale with volume: for groups of 30 or more participants, we work with multiple trainers in parallel tracks so the hands-on ratio stays high. For smaller teams of 5 to 10 people, we provide a custom approach, for example a combined session covering sales and marketing at the same time.

After the course: from literacy to implementation

A course that ends with a certificate on the final day and then gathers dust delivers no return. That is why our approach includes a follow-up programme as standard, linking the skills learned to concrete implementation projects.

Pre-course assessment

Before we start, we map out the starting level for each professional group: which tools the team already uses, where the knowledge gaps are, and which processes could benefit straight away from AI support. The assessment stops us wasting time on basics for people who are already further along, or moving too fast for those who don't yet have a ChatGPT account.

Coaching between modules

Between modules, we offer asynchronous coaching: participants ask questions in a shared workspace, get feedback on their prompts and share successes. This low-threshold follow-up makes sure the techniques they learn land in everyday workflows, rather than only working in the course environment.

Post-course assessment

Eight weeks after completion, we measure again: which tools have been integrated, which processes have been sped up, and which barriers are holding back wider adoption. The assessment feeds into the impact report and gives management hard evidence of the return on the course investment.

Impact report per department

Based on the assessments and the hands-on output, we deliver a report for each department: which processes have been identified as high-impact AI candidates, which tools deserve an organisation-wide rollout, where the risks lie and which next steps we recommend. Concrete and actionable, not a generic consultancy document.

Bridge to implementation

For clients who want to follow through on the identified impact areas, we build the implementation. Think of a RAG application on internal documentation, an agent workflow for customer service, or a copilot for finance reporting. The course provides the starting signal and the mandate; we deliver the system.

Ongoing governance support

Optional: monthly check-ins in which we update you on the AI Act, assess new tools for suitability for your organisation, and maintain the AI register. For organisations that don't want compliance to slip after the first round of the course.

GDPR, data security and using public AI

When we work hands-on with AI tools, we immediately touch on GDPR questions. Which data may go into which model? What happens to prompts in public ChatGPT? And how do you make sure your teams don't still end up pasting customer data into a free tool after the course? These questions are an integral part of every module.

What is and isn't allowed in public models

We train participants on a workable decision tree: which information is public or internal and non-sensitive, and can go into ChatGPT, Claude or Gemini? Which data is confidential or contains personal data and doesn't belong there? And what alternatives does your organisation have: Azure OpenAI with data retention controls, a locally hosted model, or an enterprise licence with tailored terms?

GDPR legal basis for AI applications

For each professional group, we cover the GDPR legal basis: what is the legal basis for processing personal data in an AI model, and how do you record it in your record of processing activities? For HR applications, we pay extra attention to profiling and automated decision-making under Article 22 GDPR.

Prompt hygiene and data leakage

The biggest practical risks lie not in model errors but in unintended data leakage via prompts. We train on concrete prompt hygiene: anonymising before sending, replacing customer names with labels, and setting up company tooling that by default makes no external API calls for sensitive data.

AI register and risk classification

For the management modules, we build a working AI register for your own organisation: which AI systems are in use, who is the deployer, which risk classification applies under the AI Act, and which measures have been taken. This document is immediately usable for audits and accountability.

Frequently asked questions about the AI course

What exactly is AI literacy, as defined in the AI Act?
The AI Act defines AI literacy as the skills, knowledge and understanding needed to deploy AI systems responsibly and to understand the potential opportunities, risks and harms. In practice, that means knowing what a model can and cannot do, how to prompt effectively, what data risks exist and when a human needs to stay in the loop. The level depends on the role: a developer needs deeper knowledge than an end user.
How long does a course take for each professional group?
We allow 4 to 8 hours of effective time per module, split into half-day or short sessions. A sales team can cover the basics in one intensive day, while engineering or management often need more time because the material is more technical or strategic. For rollout programmes across several departments, we typically plan a programme of 6 to 12 weeks.
What determines the cost of a programme?
The cost depends on the number of participants, the number of role-specific modules, the choice between in-company and online, and whether sector modules or a follow-up programme with assessments and impact reporting are included. We always prepare a tailored quote after a short intake in which we define the scope. Request a quote for your specific situation.
What is the difference between your course and a one-off ChatGPT workshop?
A one-off workshop has a random group of employees practise prompts on generic examples. Our course works per professional group, using tools and data from your own workflow, delivers AI Act-compliant certification, and ends with an impact report that lets your management team decide, on solid evidence, where AI implementation will deliver the most value. It is a programme, not a one-off training session.
Which tools are covered in the hands-on sessions?
Depending on the professional group and your existing tool stack: ChatGPT, Claude and Gemini for general use, GitHub Copilot and Cursor for engineering, Microsoft Copilot or Google Workspace AI for productivity, n8n and Make for automation, and sector-specific tools such as contract AI or agent platforms. We build on what you already use or are considering, with no marketing for a single vendor.
How do you handle GDPR and company data during the course?
In hands-on sessions we work with anonymised or fictional sample data, or with data your own team has approved for use in the course. For permanent use, we recommend tools with enterprise terms or Azure deployments where data retention can be controlled. GDPR and data security are an integral part of every module, not a separate appendix.
Do participants receive a certificate that meets Article 4 of the AI Act?
Yes. Every participant receives a certificate stating the module content, hours spent, competencies achieved and date. At organisation level, we deliver an AI literacy matrix that demonstrates the level for each role. This documentation fits into your AI register and can be used as evidence in audits, certification or incident investigations. We do not claim a legal guarantee, but we do deliver the evidence that Article 4 requires in practice.
Do you also build the AI implementations that come out of the impact report?
Yes. Appfront is primarily an AI implementation partner, and for many clients the course is the starting point of a longer programme. After the impact report, we can build RAG applications, agents, copilots, integrations or broader AI solutions where desired. You are of course under no obligation to take that programme with us; the report is yours and can be used with any implementation partner.

Ready to bring AI into your organisation in a structural way?

Schedule a no-obligation intake meeting in which we discuss your situation, professional groups and objectives. You will then receive a concrete quote with course content, schedule and optional follow-up, tailored to your organisation rather than taken off the shelf.

Request a quote

Edit content