AI Continuing Education for Employees

Within two years, AI continuing education has gone from a topic for the tech-minded to a task for personnel development. There are three reasons for this, and only one of them is legal.

The competence gap is growing faster than the tools. AI tools have arrived at almost every office workstation in a short time — in office suites, in the CRM, in translation and research tools, often without having been introduced as AI. The ability to use them sensibly has not grown along with them. The result is two groups in the same department: some save hours, others distrust the tool or rely on it blindly.

Retention and attractiveness. Continuing education in AI is currently one of the few offerings perceived as relevant across all age groups and departments. Those who offer it get conversations about development instead of about salary.

And yes, the legal situation. Since February 2, 2025, Article 4 of the EU AI Act has required fostering the AI literacy of employees. Since the Digital Omnibus Regulation of July 27, 2026, this is expressly worded as a duty to foster and no longer as a duty to guarantee. If you want to approach the topic from the compliance perspective, it is covered in detail under AI training for companies. This page takes the other view: that of personnel development.

The term is fuzzy and is often confused with “can operate ChatGPT”. For planning continuing education, it is more helpful to break it down into four abilities that can exist independently of one another:

• Contextualizing. Understanding what a language model does — namely producing plausible continuations, not determining truth. Those who have internalized this check results on their own.
• Operating. Formulating tasks so that a usable result emerges. The ability with the fastest visible effect — and the one most likely to be overestimated.
• Evaluating. Recognizing when an output is fabricated, outdated or skewed. The truly valuable competence, and the one that is trained least often.
• Taking responsibility. Knowing which data does not belong in, where a labeling obligation exists and when a decision must not be delegated to a system.

Continuing education that trains only the second ability produces employees who quickly arrive at wrong results. That is why the focus sensibly lies on contextualizing and evaluating.

They are often lumped together, and that harms both.

A mandatory training is designed for provability. It has a defined content, a completion, a date and a participant list. Its measure of success is completeness: has everyone completed it?

Continuing education is designed for impact. It may branch, offer voluntary deep dives and orient itself to the needs of the role. Its measure of success is change: do people work differently afterwards?

The practical advice: do not separate them organizationally, but do not confuse them in communication. A fundamentals training that everyone goes through and that is documented covers the compliance side. Voluntary, role-specific deep dives that nobody has to prove can build on it. Anyone who squeezes both into one mandatory measure makes it too long for compliance and too shallow for development.

In-person training. Strength: follow-up questions, discussion, trying things out together. Weakness: scheduling pressure, travel time, and with distributed teams hardly ever fully attended. The usual market cost is €500 to €2,000 per person. Makes sense for leadership circles and for roles with elevated risk.

Live online seminar. Retains the interaction, saves the travel. But the appointment remains, and with it the problem of getting everyone together at the same time.

Self-paced course. Strength: scales without additional cost, each person learns when it suits them, and completion is documented automatically. Weakness: no follow-up questions, and without a final test nobody knows whether anything got through. For the breadth of the workforce, generally the only format that realistically becomes complete.

Blended. Self-paced course for the fundamentals, in-person for the deep dive in the roles that need it. For most companies from about 50 employees, the most sensible combination.

Our offering is deliberately the self-paced course with a final test — that is, the building block that covers the breadth and takes care of the documentation at the same time. It is not intended for deep dives in specialist roles.

A single program for everyone is either too long or too superficial. Four cuts cover most organizations:

1. All employees who come into contact with AI. Fundamentals, evaluating results, data protection, internal rules. About 90 minutes, mandatory, documented. This is the basis on which everything else builds — see AI training for employees.

2. Managers. Additionally: where do risks arise in their own area, which decisions must not be delegated to a system, how to recognize AI-generated groundwork. Short, but not skippable — managers get asked and must be able to answer.

3. Subject-matter owners for AI systems. Selection and assessment of vendors, data processing agreements, risk classes, labeling obligations. In-person or a specialist seminar is worthwhile here.

4. Roles with elevated risk. Anyone who uses AI in personnel selection, lending or safety-relevant processes needs system- and task-specific training. A fundamentals continuing education does not replace that.

The range in the market is large, and the price says little about the impact. For orientation: in-person formats are usually at €500 to €2,000 per person and day, live online seminars somewhat below, self-paced courses depending on scope in the low double-digit to mid triple-digit range.

For the calculation, the price per person is less interesting than the coverage rate: a format for €800 that reaches a quarter of the workforce is more expensive than one for €15 that reaches everyone — and fulfils the duty to foster less well.

Funding. Continuing education in digitalization and AI is generally eligible for funding in Germany; the programs differ by federal state, company size and target group, and the conditions change regularly. Depending on the program, the Federal Employment Agency, the state development banks or the competent chamber are responsible. Because this is constantly in motion, we deliberately name no specific rates here — ask the body responsible for you about the current status before you budget.

Our prices: €19.90 for individuals, €14.90 per person from two, €9.90 per person from twenty — net plus VAT, one-time, no subscription. Invoice by email, for companies also purchase on invoice with a 14-day payment term.

Participation rates measure attendance, not impact. Three metrics say more:

• Completion rate instead of start rate. How many passed the test, not how many logged in. For a self-paced course, this is the most honest value available.
• Rule violations over time. If you have an AI policy, count the cases in which it was violated — before and after the measure. This number moves when the training has worked.
• Self-assessment with a control question. “How confident do you feel?” alone is worthless, because ignorance feels confident. Combined with a factual question, it becomes a usable indicator.

What you can save yourself: satisfaction surveys directly after the measure. They measure the quality of the presentation, not of the learning.

The most common course of events is a well-attended one-off measure after which nothing happens for a year and the list quietly goes out of date. Three things prevent that:

Onboarding. The fundamentals training belongs in the hiring process, not in an annual campaign. Otherwise coverage is patchy again after six months without anyone noticing.

A fixed repetition rhythm. Not because a law requires it, but because tools and the legal situation change faster in this field than in most others — the wording of Article 4 changed in 2026 alone. A short annual refresher part is enough.

A policy that the training refers to. Knowledge without a rule changes little. A rule without knowledge gets circumvented. Both together work.

If you are looking for an entry point via the professional field, you will find adapted case examples for tax advisors, law firms, care facilities, trade businesses and municipalities. If you think from the tool: ChatGPT training. And if you want to know what the format and process look like: AI course online.

We offer the building block that covers the breadth: an online self-paced course with a final test, completable in 60 to 90 minutes, with 46 learning cards, 18 exam questions from real work situations and five optional additional modules. Each person finishes with a dated certificate of participation that can be verified via QR code. In the team area, you see who is still missing.

What it is not made for: it does not replace role-specific deep dives for subject-matter owners or system- and task-specific training for high-risk applications. It is also not an accredited certificate — a certification body provided for by the EU AI Act does not exist, and anyone who advertises with one claims more than exists. What you get is a cleanly documented fundamentals proof for all employees, at a price at which completeness is realistic.

View prices and order · Try chapter 1 for free

In common usage, “AI Fortbildung” (professional development) and “AI Weiterbildung” (continuing education) are mostly used synonymously — both mean building competence in dealing with AI systems. Formally, there is a subtle difference: Fortbildung deepens knowledge in the profession already practiced, Weiterbildung builds more broadly based competencies, also beyond the current area of responsibility.

For AI literacy under Article 4 of the EU AI Act, this distinction plays no great role in practice. What is decisive are the content and documentation of the measure, not the label under which it is run internally.

Our online course covers both perspectives: it deepens understanding for the concrete everyday work with AI tools and at the same time builds a basic understanding that extends beyond the individual use case.