The AI Act's Article 4, mandating AI training for employees, takes effect in February 2025. This requirement perfectly aligns with what separates companies that profit from AI from those that merely pay for licenses: well-trained teams who use tools judiciously. Fulfilling this mandate isn't an expense; it's the highest-ROI investment available, often covered by existing training credits. Only 21% of Spanish companies with over ten employees reported using AI in 2024, according to ONTSI and INE, indicating a massive competitive advantage for early adopters of AI training.
For months, you've heard that regulations require you to train your team. There's talk of fines, inspections, and compliance—as if training were a punishment. I want to flip that narrative because it's precisely the opposite.
What Happens When a Team Truly Understands How to Use AI?
There are two types of companies using artificial intelligence tools. The first type bought access to ChatGPT, Copilot, or similar platforms. Some employees use them, others don't. No one quite knows what can or cannot be done with them. The result: enthusiastic adoption without discernment, mixed with unfounded distrust.
The second type trained their teams before distributing access. They explained what these tools do, where they fall short, and how to leverage them. The outcome is radically different.
I've seen operations teams halve the time spent processing documents. Legal departments generate drafts in minutes instead of hours. Customer service teams resolve most frequent inquiries with automated workflows. Finance departments produce weekly reports in a third of the time. None of this happened because they bought a better tool. It happened because someone trained the people using it.
The difference between these two types of companies isn't the technology. It's the judgment of the people operating it.
How Much Money Is Left on the Table Without AI Training?
I'll be specific, because abstract arguments about digital transformation don't help anyone.
A company of twenty people, where five dedicate two hours daily to tasks that AI could significantly accelerate, is losing the equivalent of ten hours of work per day. With average salaries of thirty thousand euros, that's over twenty thousand euros annually paid for work that could be done in a fraction of the time.
I'm not inventing this scenario; I'm describing it. The report that takes four hours because no one knows how to ask AI to structure it. The analysis that requires two days because the team doesn't know how to connect the tool with their data. The executive summary manually written by someone whose time is far more valuable.
That's what's being left on the table. And that's what real training makes accessible to any team.
The Gap Between Companies That Train and Those That Don't Is Already Visible
According to 2024 data from ONTSI and INE, only 21% of Spanish companies with over ten employees reported using artificial intelligence in their operations. Nearly eighty percent of the business fabric is missing out on something their competitors, both within and outside Spain, are already using to be faster, cheaper, and more precise.
Training is precisely what separates companies that successfully adopt AI from those that buy it but fail to integrate it. Having access to a tool without understanding it isn't a competitive advantage; it's noise. And noise consumes time instead of generating it.
The companies gaining ground aren't those with the most expensive software. They are those with teams that understand what they're doing: this is exactly what such a team needs to know.
Why the Legal Obligation Is Late, Not Early
What's revealing, and I say this without irony, is that the obligation under Article 4 has been in effect since February 2025, and most companies still haven't complied. Not by choice, but because they don't know it exists, or they treat it as a task for later. What exactly it requires, here.
But if you look at the speed at which AI is changing work processes, the obligation isn't coming soon. It's coming late. Companies that started training in 2023 have a two-year head start. Those that do it now have room to catch up. Those that keep postponing will see the gap widen.
Complying with Article 4 is not the goal. It's the minimum. The real goal is a team that uses AI judiciously enough to make the company more competitive next year than this year.
And cost is no excuse: this training is eligible for FUNDAE credits your company already has allocated. In many cases, the additional cost is zero. The law not only tells you to do it but also provides the mechanism to do so at no cost. At 4Geeks, we apply this with the AI Fluency program, designed for non-technical teams.
So, when someone tells you that the European AI Act requires you to do something, don't see it as a burden. See it for what it is: an official reason to do something you should have already done.
The company next door is thinking the same thing. The difference is who acts first.
Víctor Gómez is CEO and co-founder of 4Geeks Academy Spain.