The Unspoken Problem with AI Training in Companies

Picture this. A training manager completes an online course on artificial intelligence. An hour and a half. Certificate included. They go back to their desk, open ChatGPT, and sit there staring at the screen with no idea what to ask it or how.

This scene is playing out in thousands of organisations. And the data confirms it.

That is the finding of the study on AI in HR and L&D that we produced alongside Microsoft, with the participation of more than 300 professionals from 32 countries.

The takeaway? More than half of organisations are investing in AI training that simply isn’t working. And the most striking thing is that almost nobody is saying it out loud.

The problem isn’t the technology. It’s the training about the technology

When you analyse why real AI adoption remains so limited —only 23% of the market has achieved genuine integration into its processes— the answer that emerges isn’t a lack of tools or budget. It’s a lack of applied knowledge.

And here lies the paradox the study uncovers: companies are training, but they’re not transforming.

Why? Because AI training being rolled out in most organisations today has three structural problems:

• It’s too generic: it explains what a language model is, how ChatGPT works, why AI is going to change the world. All very interesting. Nothing applicable on Monday morning.
• It’s too theoretical: it stays at the conceptual level and never reaches real practice — how to build a prompt that works, how to iterate, how to validate that the output is good enough to use.
• It’s too fragmented: it reaches some team members, not all. It creates islands of knowledge within organisations rather than collective capability.

The result is predictable: professionals who have “done an AI course” but still don’t know how to integrate it into their real workflow. The course happened. The change did not.

The data point that explains everything

The study reveals one of the most compelling correlations in the entire research: professionals with a high level of AI knowledge have an advanced adoption rate of 76%. Those with a low level: just 6%.

There is no shortcut. There is no way to reach genuine integration without going through genuine knowledge. And genuine knowledge doesn’t come from a ninety-minute webinar on the future of artificial intelligence.

What the study also shows is that when training works, the results are dramatically different. Organisations that offer mandatory, practical training have achieved real integration in 39% of cases, compared to 12% among those that offer no training at all. Three times more. Simply by doing well what others do poorly.

The irony nobody wants to name

There is something deeply ironic about all this that deserves to be said. L&D and HR departments have spent decades designing learning experiences for others. They know better than anyone the difference between training that transforms and training that merely logs hours. They know that learning without practical application is forgotten within 72 hours. They know that role-based personalisation multiplies transfer to the job.

And yet, when it comes to training themselves in AI, they are receiving exactly the kind of training they would never design for their own employees: Generic. Theoretical. Fragmented.

What AI training needs to actually work

The study data is quite clear about what distinguishes training that moves the needle from training that doesn’t:

• It must be role-specific. Using AI in a sales team is not the same as in an operations or HR department. The tools are the same; the use cases are not.
• It must be practical from the very first module. Not “what is AI” but “how do I use AI to do what I do every day, better and faster”.
• It must integrate into the organisation’s real workflows, not exist as a training island disconnected from day-to-day work.
• And it must reach every member of the team, not only the most curious or those who already have a tech inclination.

That is precisely what isEazy has built together with Microsoft in the Microsoft Copilot School: a training catalogue designed specifically so that teams learn to apply Copilot AI in their organisation’s strategic processes. With general courses, specialist modules by tool — Word, Excel, PowerPoint, Outlook, Teams — and learning paths adapted by professional profile: HR, Marketing, Sales, Operations, Finance.

Practical training, with real use cases, endorsed by Microsoft and with dual accreditation. On one hand, official Microsoft diplomas and badges issued through Microsoft Learn. On the other, isEazy Skills certificates that can be shared directly on LinkedIn.

Because the difference between knowing Copilot exists and knowing how to use it to transform the way your team works is exactly the same as the difference between that 6% and that 76% the data talks about. Which side is your team on?