The technology is ready. Is your team? Why the value of AI starts with employee skills
AI skills for employees decide what AI delivers. Why people, not tools, are the bottleneck, and how to build skills that change the way work gets done.
Only one in three employees say they have been properly trained to use AI, according to a 2025 BCG survey of more than 10,600 workers in 11 countries. That is the real bottleneck. The tools can already do far more than most companies use them for; what is missing is the skill to fit them into daily work. AI skills for employees are not about knowing which buttons to press. They are about knowing which tasks to hand over, how to judge the output and how to change the way a process runs. Companies that invest in that, with training and with time, are the ones that see AI pay off.
Why does AI so often disappoint, even when the technology works?
Because value comes from changing the work, and that takes people. In AG Connect, data and AI adviser Maarten van den Outenaar compares AI with earlier technology waves in his book “Geen woorden maar data” and concludes that the key to value lies with employees. History supports that view.
When factories switched to electricity, productivity did not rise for decades. Economist Paul David showed in his classic paper “The Dynamo and the Computer” (American Economic Review, 1990) that the gains only came in the 1920s, around forty years after the first power stations. They arrived once factories were redesigned around the new technology and engineers had learned, through practice, how to work with it. The motor was ready long before the people and the processes were.
We think AI is following the same pattern, only faster. The technology runs ahead. The work around it, and the skills to change that work, run behind.
What is the gap between using AI and getting value from it?
Many people already use AI, but mostly on their own. In 2025, 43 percent of Dutch people in paid work used AI in their job, according to Statistics Netherlands (CBS). The spread is uneven: 55.5 percent among people with a higher professional or university education, against 19.7 percent at the lowest education level.
Training lags behind that use. In the 2024 Work Trend Index of Microsoft and LinkedIn, only 39 percent of AI users had received AI training from their employer. People teach themselves, which is fine for writing a faster email, but it rarely changes how a team works.
In our view there is a clear gap between being interested in AI and actually changing something. A good session creates energy, and then customers, colleagues and the daily rush take over again. Smaller companies feel this most, because they often lack time, people and means to follow up. That is not a lack of willingness. It is a capacity problem, and it deserves to be planned like one.
Which AI skills do employees actually need?
Fewer prompt tricks, more judgement. AI changes tasks more than it changes jobs. The work moves from doing routine tasks yourself to checking and critically assessing what AI produces. That makes three skills central.
Knowing what to hand over. Recognising which tasks in your own work are repetitive, rule-based and digital, and which need your judgement or a conversation with a client.
Judging the output. Spotting a wrong figure, a missing nuance or a confident answer that is simply incorrect. Human expertise and quality control still matter; they just move to a different place in the process.
Changing the process, not just the task. Asking: if this step takes five minutes instead of an hour, what should the rest of the process look like? This is where individual efficiency turns into real value, and it is the skill that is trained the least.
There is also a legal floor. Under Article 4 of the EU AI Act, companies that use AI must take measures for the AI literacy (AI-geletterdheid) of their staff. The European Commission’s Q&A confirms that this has applied since 2 February 2025 and remains an obligation after the Digital Omnibus of 2026. No certificate is needed. This is not legal advice.
How do you build AI skills that actually stick?
By making it hands-on, tied to real work and supported over time. BCG found that regular use is sharply higher among employees who received at least five hours of training plus in-person training or coaching. Spanish bank BBVA started in 2024 with 3,300 ChatGPT licences, combined with training and an internal community where staff share what works; the bank reports that 83 percent use it daily, employees save around 2.8 hours a week and staff built more than 3,000 assistants of their own. Closer to home, Amsterdam’s public transport company GVB describes in Berenschot’s AI trend survey 2026 that colleagues bringing colleagues along works best, and that neither top-down nor bottom-up works on its own.
From those examples and our own view, a few principles follow.
Open the laptops. Inspiration sessions are pleasant, but skills grow when people work on their own tasks with the tool in front of them.
Start from the organisation, not from a tool. Ask where you are now, where you want to be in a few years, and which tasks take the most time or are liked the least. Then choose the tool.
Pick one application and involve the right people. The employees who do the work every day know where it gets stuck. Let them experiment, learn and show their colleagues.
Budget time, not just training. Building skills asks for training and for extra time from employees. That is a deliberate investment, and it is the one most often forgotten.
From tools to a way of working
A licence does not change how work gets done; people do. The companies that get value from AI treat skills as part of the change, not as a course at the end. If you want help shaping that, have a look at how we help SMEs put AI to work or at the AI training and coaching from our experts, who combine training with work on your own processes.
Frequently asked questions
What AI skills do employees need?
Mainly three: recognising which tasks AI can take over, critically judging AI output, and rethinking the process around the task. Knowing a specific tool helps, but tools change quickly; judgement and process thinking last.
How much AI training do employees need?
There is no fixed number, but BCG found in 2025 that regular use is much higher among employees with at least five hours of training plus in-person training or coaching. Short, hands-on sessions on real tasks, repeated over time, tend to work better than one long course.
Is AI training mandatory for companies?
Under Article 4 of the EU AI Act, companies that use AI must take measures for the AI literacy of their staff. According to the European Commission, this has applied since 2 February 2025 and remains an obligation after the 2026 Digital Omnibus. A certificate is not required; keeping a record of what you did is enough.
Why don’t AI tools deliver the expected results?
Often because the tool was added to the existing way of working without changing the process or the skills around it. History shows the same pattern: electricity only raised factory productivity once factories and work were redesigned around it.
Your next step
Pick one task your team does every week and ask the people who do it where AI could help. That conversation is the start of building skills. When you want support with training or a first application, post your question on Baionic: within one working day you receive proposals from one to three matching experts, with their approach, price and planning side by side.
Ready to take the next step?
Post your project for free and receive proposals within one business day.