How an AI startup built AI literacy under Art. 4 EU AI Act the lean way

Starting point: an AI-native team meets a new obligation

After a large funding round, an AI startup grew quickly – new roles, new teams, a product that is fundamentally built on artificial intelligence. Along with the growth came an obligation that can easily be overlooked in the hectic scaling phase: proof of the AI literacy of the workforce under Article 4 of the EU AI Act (Art. 4 Regulation (EU) 2024/1689). Since February 2, 2025, every organization that uses AI systems has had to take measures to support the development of the AI literacy of its staff (according to the version under Regulation (EU) 2026/1744 applicable since July 27, 2026; previously: to ensure a sufficient level of AI literacy) – regardless of company size or maturity, and regardless of whether an organization acts as a provider of its own AI systems, as a deployer of third-party AI systems in everyday work, or – as with many AI start-ups – in a dual role. For a company whose entire business model is based on AI, this is not a peripheral topic for the legal department but directly affects the credibility of its own product with customers and investors. The project described still ran under the original version of Article 4 (editorial status: July 3, 2026).

The special feature of this case: The team had long been actively using AI in everyday work to build its own product. A classic introductory training course starting with the basics would have missed the reality of such a team – and would have cost valuable time without delivering any real gain in knowledge. What was sought was therefore not an elaborate internal training project with its own concept development and content creation, but a solution that provides the formal proof without tying up development time or product focus.

What Article 4 of the EU AI Act specifically requires

Art. 3 no. 56 EU AI Act defines AI literacy as the skills, knowledge, and understanding that allow providers, deployers, and affected persons to make an informed deployment of AI systems and at the same time to be aware of the opportunities, risks, and possible harm of AI use. Crucial for practice: Art. 4 does not prescribe a uniform catalog of obligations but requires measures that are guided by the technical knowledge, experience, and education of the respective staff as well as by the specific context of use. So there is no "one training course for all" template – the appropriate depth depends on the team's actual level of maturity, and in an AI-native company this level of maturity can be very high technically but still patchy from a regulatory perspective.

For a startup whose employees already use AI tools professionally every day, this means: The formal proof must be provided, but conveying basic technical knowledge is not the core of the task. The actual need lies elsewhere – in the regulatory translation of what the team already masters in practice.

The blind spot of AI-savvy teams

A team that routinely uses AI tools for its own product development does not automatically have the regulatory perspective that Art. 4 requires. Practical application experience – for example in the secure handling of models, prompts, or interfaces – is something different from the knowledge of obligations, risk classification, and limits that the EU AI Act understands by "AI literacy". Those who work with AI every day can easily consider themselves sufficiently informed, even though technical application competence and regulatory understanding are two different things.

This was precisely the real challenge of the case: Training that starts with "What is AI?" would have under-challenged the team and been perceived as a mere compulsory exercise – with the risk that the content would only be absorbed superficially. Conversely, training that ignores this difference and assumes that practical experience automatically includes regulatory knowledge would have left open precisely the gap that Art. 4 is actually meant to close.

Implementation: a compact training course instead of a basics marathon

Adapting the format to the level of maturity

A note on our own behalf: Bridgly provided a single, focused AI training course that addresses exactly this point: It conveys the regulatory basics under the EU AI Act compactly and concentrates on the obligations, risks, and limits of AI use in companies instead of basic knowledge that already existed in the team. Provided via the Bridgly learning environment, the training was ready to start immediately – as a fully rolled-out format, not as an internal project with its own design and coordination effort. Employees completed it flexibly at their own pace, without IT setup and without friction with the existing toolchain.

Proof built in from the start

So that the effort remained lean not only in terms of content but also organizationally, participation ran via the Bridgly learning platform: Assignment, progress status, and completion are automatically documented there for each person. For a growing team, this means that proof of AI literacy does not have to be laboriously compiled after the fact but is available in an orderly manner from the start – a prerequisite for being able to repeat the training for new team members when needed without additional organizational effort.

Result: proof provided, focus on the core business

The startup trained its staff in AI literacy under Article 4 of the EU AI Act and documented their participation without having to set up its own training project or tie up development capacity. Proof of AI literacy is documented and can be provided when needed – for example to investors, business partners, or as part of an internal audit. The fact that the compact approach also worked with an AI-experienced audience was due to its consistent focus on the regulatory perspective instead of technical basics that were already known: The training was perceived as a relevant addition to the team's own knowledge, not as a waste of time alongside day-to-day business.

Conclusion: three principles for maturity-appropriate AI training

• Link the format to the level of knowledge. An AI-savvy team does not need basics but precision regarding obligations, risks, and limits – anything else only costs acceptance and time.
• Application knowledge is not the same as regulatory knowledge. Those who use AI productively every day do not automatically know the obligations under Art. 4 EU AI Act – this difference belongs explicitly in the training.
• Think about proof from the start. A learning platform that logs assignment, progress status, and completion makes proof of AI literacy retrievable at any time – even as the team grows.

Further Bridgly case studies from public administration, SMEs, and large corporations show how other organizations implement their mandatory training with Bridgly.

Do startups have to train their employees in AI?

Yes, as soon as they use AI systems. Since February 2, 2025, Art. 4 of the EU AI Act has obliged all organizations – regardless of size – to take measures to support the development of the AI literacy of their staff (version since July 27, 2026). Startups are covered too.

How do you train a team that already uses AI every day?

Not with basics, but with a compact training course that focuses on obligations, risks, and limits. This keeps the training relevant, and it is not perceived as a waste of time.

How extensive is AI training under Art. 4?

The scope depends on role and risk profile. For a team that already uses AI confidently, a compact training course is sufficient that focuses on the regulatory obligations and limits and is completed flexibly at your own pace.

How do you prove AI literacy under Art. 4?

Through documented participation with traceable knowledge transfer. A learning platform that logs assignment, completion, and date for each person makes the documented proof retrievable at any time.

• Regulation (EU) 2024/1689 (EU AI Act), Art. 4 – AI literacy, EUR-Lex – accessed 30 Jun 2026.
• Regulation (EU) 2024/1689 (EU AI Act), Art. 3 no. 56 – definition of "AI literacy", EUR-Lex – accessed 30 Jun 2026.
• Regulation (EU) 2026/1744 ("Digital Omnibus"), Art. 1 no. 5 – new version of Art. 4 EU AI Act, in force since 27 Jul 2026: EUR-Lex – accessed 4 Sep 2026.
• Featured image: AI-generated; labeled "Created with AI" as an overlay in line with Art. 50 EU AI Act. Proof of license/creation will be added during image production (provider/date).