Made in Germany · ISO 27001 · GDPR-compliant
Mandatory AI Competency Under the EU AI Act
Here's How L&D Teams Implement Article 4 in a Way That Can Be Audited
Artificial intelligence has become part of everyday work life: in learning platforms, HR processes, and digital assistance systems. Article 4 of the EU AI Act (Regulation 2024/1689) has required companies to demonstrate so-called “AI literacy”—verifiable AI competence—since February 2, 2025. Many organizations face the same structural problem: AI is being used, but AI competence is not organized in a verifiable manner.
Three key questions arise for L&D and HR teams:
• What exactly does Article 4 of the EU AI Act require—and when does this obligation take effect?
• Which employees must demonstrate AI competency?
• How do we document training so that it passes an audit?
This page shows how companies can implement the AI competency requirement in a structured way and how SoftDeCC turns this into a seamless, audit-proof process.
Burden of Proof
Art. 4 of the EU AI Act
AI Competence
What Article 4 of the EU AI Act Actually Means
Article 4 requires companies to ensure that individuals who use AI systems or interpret their results possess sufficient AI literacy. This requirement applies regardless of company size or industry and has been in effect since February 2, 2025.
The EU AI Act does not prescribe any specific course formats. However, it requires verifiable, role-specific qualifications—and these must be fully documented in the event of an audit.
| Role | AI Competency Requirements |
|---|---|
| Employees | Basic Understanding of AI Use and Risks |
| Managers | Evaluating AI Decisions in the Workplace |
| HR / Recruiting | Using AI in HR Processes |
| L&D / Academy | Development and Documentation of Training Structures |
| IT / Data Teams | Governance, Control, Data Risks |
Challenge
The Real Problem in Companies
Many organizations have already implemented AI training, but:
❌
Evidence is scattered across various sources (PDF, Excel, LMS, email)
❌
No uniform competency framework exists
❌
No recertification structure has been established
❌
Audit readiness has not been established
❌
No relation to roles or risk classes
Result: Training provided, but compliance cannot be verified.
Added Value: Interfaces and Data Exchange
How AI Competence Must Be Demonstrated in a Way That Can Be Audited
The EU AI Act does not define specific course formats—but it requires traceability, which must be demonstrated in the event of a dispute.
Minimum Requirements for Auditable AI Compliance
| Type of Evidence | Meaning |
|---|---|
| Training Assignment | Who Needs to Learn What |
| Learning Path Structure | What Content Applies to Each Role |
| Completion Status | Who Has Completed What |
| Competency Test | Demonstration of Understanding |
| Certificates | Formal Documentation |
| Recertification | Current Level of Competence |
| Audit Logs | Full Traceability |
Why traditional solutions aren’t enough: Excel spreadsheets aren’t audit-proof, one-on-one training isn’t scalable, PDF certificates aren’t auditable, and manual reminders are prone to errors. AI-driven compliance doesn’t fail because of the training—it fails because of the structure of the documentation.
References
What Our Customers Say
TCmanager LMS offers everything you would expect from a software tool for training centers. Today, our extensive training offerings for our service and sales partners and customers are optimally supported worldwide by training portals tailored to specific target groups and countries. The management of schedules, bookings, and resources is logically structured and intuitive to use. Reports can be generated quickly and clearly at any time. SoftDeCC is our reliable partner, with whom we will continue to deliver our in-person training and e-learning offerings in the future.
YASKAWA has been a customer and user of TCmanager® since 2015. After an extensive search for a management tool for our training and continuing education division, we are now glad—following its implementation—that we chose SoftDeCC. It quickly becomes clear that this software was specifically designed for training and qualification management. We’re discovering far more capabilities than we initially required or expected. We’re confident that TCmanager® will help us position our training programs well for the future and improve our efficiency and quality.
SoftDeCC did an excellent job of bringing to life the high standards we set for functionality on the one hand and user-friendliness on the other through customized solutions. The expert advice and support, as well as the constructive and solution-oriented discussions, were particularly helpful in this regard. The collaboration has been excellent, thanks to the very pleasant working relationship and high level of reliability.
For more than 30 years, training and education for Schneider Electric’s customers and partners have been closely linked to quality and reliability. In 2019, we chose SoftDeCC’s TCmanager® LMS (Learning Management System). The support we received from SoftDeCC during the implementation phase was extremely helpful. Since then, the learning system’s feature set has been continuously expanded, offering more options to meet our growing needs. The SoftDeCC team is very familiar with our application and has been a great help!
Hella Academy provides technical training to repair shops and retail partners on our products, as well as on specialized topics such as “Hybrid & High-Voltage” and driver assistance systems. We have been a satisfied customer and user of theTCmanager® LMS since 2015. We are particularly impressed by the support time and again. Any customizations or new features we request for the application’s functionality are implemented smoothly and help us better serve our customers.
Practical Implementation
Here's How SoftDeCC Implements the AI Competency Requirement in Practice
Why traditional solutions aren’t enough: Excel spreadsheets aren’t audit-proof, one-on-one training isn’t scalable, PDF certificates aren’t auditable, and manual reminders are prone to errors. AI-driven compliance doesn’t fail because of the training—it fails because of the structure of the documentation.
Role-Based AI Competency Models
AI proficiency is structured according to risk and role: basic knowledge for all employees, advanced modules for managers, compliance training for HR, and governance modules for IT. Each role is automatically assigned the appropriate learning path—without the need for manual, individual assignments.
The Process Within the System – Example: New AI Policy
HR defines AI competency requirements → SoftDeCC automatically assigns learning paths to roles → Employees receive mandatory training → Completion is documented → Competency status is updated → Recertification
Auditability & Documentation
SoftDeCC ensures that AI compliance remains verifiable: complete audit logs of all training activities, exportable compliance reports, time-stamped certificates, role-based traceability, and recertification cycles.
AI Compliance Is Becoming Scalable
Particularly relevant for larger organizations: centralized management of all AI training programs, automatic updates when new requirements arise, consistent governance across all locations, and a reduced manual HR workload.
Data Protection and Employee Participation
GDPR, Works Council, and AI Governance
AI literacy is not just a matter of learning, but also a matter of data protection and employee participation.
Section 87 of the Works Constitution Act (BetrVG): When the Works Council Must Be Involved
As soon as AI systems or learning analytics evaluate behavior, derive performance indicators, or influence HR decisions, the right to co-determination under Section 87 of the Works Constitution Act (BetrVG) applies in Germany. This applies in particular to AI-supported learning and assessment systems in a corporate context.
Governance Structure for Companies: Take inventory of AI systems → Define risk categories → Determine competency requirements → Define training logic → Centralize documentation → Establish audit processes → Involve the works council and data protection officers.
Action Plan for &D and HR Teams:
| Step | Objective |
|---|---|
| AI Systems Capture Data | Create Transparency |
| Define Roles | Structure Competency Requirements |
| Set Up Learning Paths | Standardize Training |
| Centralize documentation | Ensure audit readiness |
| Plan for Recertification | Ensure Up-to-Date Status |
| Establish Governance | Ensure Compliance |
Free Consultation
Is your company ready for Article 4 of the EU AI Act?
Next Steps:
• Schedule a demo
• Review integrations
• Discuss industry use cases
✓
Over 25 Years of Expertise
✓
Made in Germany
Frequently Asked Questions
FAQs on the AI Competency Requirement
What exactly does Article 4 of the EU AI Act require of companies—and since when?
Article 4 of the EU AI Act (Regulation 2024/1689) requires companies to demonstrate verifiable AI literacy and has been in effect since February 2, 2025—not since the regulation took effect in August 2024. This applies to all individuals who use AI systems or evaluate their results. The law does not prescribe a specific course format, but it does require traceability.
Is a one-time AI training session sufficient for compliance with the EU AI Act?
No. Article 4 requires that AI proficiency be kept up to date and verified. One-time training sessions without recertification, without a documented certificate of completion, and without a role-specific focus do not meet this requirement—neither formally nor in the event of an audit.
Who in the company is subject to the AI competency requirement?
All individuals who use AI systems or evaluate their results from a technical perspective: employees, managers, HR, L&D teams, and IT governance—each with a different scope of expertise depending on the risk and decision-making context.
What distinguishes audit-ready AI compliance from standard training certificates?
Audit-ready AI compliance requires more than just PDFs or Excel spreadsheets: documented training assignments, competency assessments, time-stamped certificates, recertification cycles, and complete audit logs.
What role does an LMS play in complying with Article 4 of the EU AI Act?
An LMS serves as the operational foundation: role-based learning path assignment, completion tracking, certificate issuance, automated recertification, and exportable compliance reports—all centralized within a GDPR-compliant infrastructure.
When should the works council be involved in AI training?
As soon as AI systems evaluate behavior, derive performance indicators, or influence HR decisions, Section 87 of the Works Constitution Act (BetrVG) applies. L&D teams must involve the works council at an early stage when introducing AI-supported learning and assessment systems.
