The Team-Training-App turns the EU AI Act AI-literacy obligation into a documented, tamper-proof process - on your server or ours.
Article 4 of the AI Regulation (EU) 2024/1689 obliges every company that uses AI systems - from corporations to small businesses with ChatGPT access - to take measures that foster their employees' AI literacy. What counts in the end is what you have actually done and can evidence. On August 2, 2026, the transparency obligation under Art. 50 also entered into force.
Since this date, companies must take measures that foster the AI literacy of their employees. This obligation applies now - not at some point in the future.
This frequently cited deadline concerns labelling and transparency obligations - not the AI-literacy obligation. Anyone waiting until then is already more than a year late.
In a dispute, what counts is what you can prove: who was trained when, with what content, with what result. That is exactly what the Team-Training-App delivers.
Common misconception: The two deadlines are constantly confused. The AI literacy obligation (Art. 4) has applied since February 2, 2025 - regardless of which AI systems you use. It is not a future task but an already ongoing obligation. In addition, the transparency obligation under Art. 50 entered into force on August 2, 2026, and non-compliance can now be sanctioned by the relevant authorities.
What the AI Act does not require: There is no obligation to use external training, no obligation to obtain external certification and no prescribed training format. Art. 4 requires appropriate measures matched to role, prior knowledge and the systems actually in use, and, in case of doubt, evidence that those measures were taken. Internal documentation can serve as that evidence too. The Team-Training-App is one way to keep that evidence robustly and without effort on your side, not a legally prescribed one.
Putting the fines in context: The AI Act provides no separate fine for the AI literacy obligation itself. The frequently quoted 35 million euros or 7 per cent of worldwide annual turnover relate to the prohibited AI practices under Art. 5. Ignoring Art. 4 therefore does not put that amount at risk, it puts your evidence at risk: when a market surveillance authority asks, during a customer's supplier review, after an incident involving an AI tool and towards your own insurer. Who was trained when and on what then has to be on the table.
Last reviewed: July 2026, including the amendments introduced by the Digital Omnibus (Regulation (EU) 2026/1744).
Do you have sufficient AI literacy? Do you know Art. 4 and Art. 50 of the EU AI Act? Test your knowledge now!
The Team-Training-App is a standalone, multi-tenant training software by Klivantis - installed on your server or hosted by us. Customise the app and certificates by adding your own company logo or adapting the colours to your corporate identity. Real screenshots, no mockups:

Dashboard & course overview - each role sees exactly the courses approved for it.

Course creation - text, images, video & audio via WYSIWYG, no technical skills needed.

Tamper-proof certificate - verifiable via QR code at any time.
Your admin sets the default language in the backend - every employee can switch individually.
From the course editor to deletion with instant anonymisation - built for companies that take compliance seriously but do not want to dedicate an IT department to it.
Self-employed professionals, freelancers and small teams can also complete the courses on our own platform.
Your own trainings via WYSIWYG editor, no technical skills needed. For AI literacy - and any other training in your company.
Every certificate carries a unique ID and can be verified against your installation via QR code at any time - an invented credential is exposed immediately.
Buy ready-made Klivantis courses, import them in a few clicks and update them when the law changes - progress and certificates stay intact.
Five roles with role-based course visibility - from admin to employee.
Security updates, bug fixes and new features - provided and signed by Klivantis.
The app stores only necessary user data and is built for verifiability - including deletion with instant anonymisation.
Affordable even for small and medium-sized businesses - Klivantis quotes prices in the free consultation, tailored to size and needs.
You do not have to create content yourself: Klivantis delivers ready-to-use courses for purchase that you import directly into your app - or complete on our platform. For the EmpCo Directive on environmental advertising (not an AI topic), see our topic page empco.klivantis.ai.

What the AI Act demands of every company, which deadlines apply and what Art. 4 and Art. 50 mean concretely for everyday work.

Practical AI literacy: writing effective prompts, recognising limits, critically reviewing results.

Using generative AI productively and responsibly in daily business - from text to imagery.

Understand the differences between chatbots, AI assistants and AI agents, and put the right system to work in daily business.

The advanced follow-up to "Prompting for Professionals": chain-of-thought, tree-of-thoughts, multi-persona prompting and prompt chaining with real-world case studies.

How a language model turns tokens and probabilities into text, why it can hallucinate - and how the same technical foundation powers chatbot, assistant and agent alike.

The technical follow-up to "From Chatbot to Agent": ReAct, plan-and-execute, reflection, memory engineering and multi-agent orchestration with real-world case studies.

The follow-up to "Autonomous AI Agents": where an agent gets reliable facts. From the RAG pipeline through the limits of vector search to knowledge graphs and GraphRAG.

Where AI helps decide in recruiting and during employment, when a tool falls into the high-risk category, and what oversight and codetermination look like in practice.

Labelling duties under Art. 50, image tools such as Canva and Photoshop, rights in AI-generated content, and the line to misleading advertising.

Traceability and retention, creditworthiness assessment as a high-risk case, figures from language models, and protection against forged payment instructions.

What comes with the leadership role: deciding on tools, resourcing oversight, governing meeting recordings, and responding properly to incidents.

When customers must be told they are talking to a machine, why measuring emotions is banned for your own team, and who answers for a false promise.

Coding assistants in daily work, the shift from deployer to provider, prompt injection, and agents with access to live systems.

The AI question in every purchase, vendor checks with the right evidence, contract clauses, and the special case of supplier scoring.

Other people's data in other people's tools, deadlines in an auto-sorted inbox, minutes and recordings, and what side copies mean for deletion periods.
Free and without obligation - we show you the Team-Training-App live, e.g. via online meeting, and answer all questions about plans and prices.
We use cookies: technically necessary ones always, everything else only with your consent. Nothing is loaded or tracked before you decide. Details in our privacy policy.
Language choice, theme, consent storage - the site does not work without these.
Anonymous usage statistics. We currently do not use any analytics services.
Third-party content (e.g. external videos or fonts). Currently not in use - everything is self-hosted.