Based on reporting by El Chapuzas Informático (February 2026)
For years, we have been told that artificial intelligence would transform the way we work. Faster results, smaller teams, unbeatable efficiency... The promises were ambitious, and investment followed right behind. However, according to a surprising new survey by the National Bureau of Economic Research, the reality on the ground tells a rather different story.
The study, which involved more than 6,000 CFOs and senior executives from companies across Europe and the United States, revealed that more than 80% of companies using generative AI have not experienced a significant improvement in productivity. Let's think about that for a moment. The vast majority of companies that have actively integrated AI into their workflows are not getting the benefits they were promised.
Speed isn't everything
It's easy to assume that since AI can generate content at a pace no human could match, productivity should naturally follow. But speed is only part of the equation. If the output consistently requires extensive review, data verification, or structural rewrite, the time saved in creation is often lost in correction. Many professionals have quietly come to this conclusion on their own: AI can draft a paragraph in seconds, but turning it into something truly useful can take as long as writing it from scratch.
This subtle point tends to get lost in the broader debate about AI, which has often fluctuated between wild enthusiasm and total rejection. When generative AI first entered the public eye, sceptics were quick to dismiss it as a passing fashion. Those voices have largely been silenced, as adoption has grown steadily year after year. But now, the pendulum may be swinging in a more measured direction: not rejection, but honest re-evaluation.
Adoption is high, impact is low
The survey reveals that around 70% of companies already use AI on a regular basis, with newer companies particularly keen to adopt it. Despite this wide adoption, over 80% of those companies say it has not had a significant impact on productivity over the past three years. The rest reported modest and limited improvements, a far cry from the transformative leap the industry has been heralding.
Interestingly, more than 66% of senior executives say they use AI tools on a regular basis, although the actual time they spend on them is surprisingly low: only about 1.5 hours per week on average. This suggests that, at least at the management level, AI is used for very specific and concrete tasks, rather than becoming a fundamental part of daily decision-making.
A bubble that has not yet burst
There is a growing sense that the AI sector is caught in a peculiar bubble, with investment continuing to pour in at an extraordinary rate, while profits remain hard to come by. Companies are betting heavily on a technology they do not yet know how to use effectively, and employees are being pressured to adopt tools that may not be truly useful for their specific roles. In some sectors, such as video game development, the pressure to integrate AI has met with significant resistance from workers, who feel it is undermining their craft rather than supporting it.
None of this means that AI is useless or that it will disappear. Quite the contrary: it is here to stay, and dismissing it entirely would be as short-sighted as blindly defending it. But the data suggests that simply implementing AI is not, in itself, a strategy.
What comes next?
Despite disappointing results so far, most companies remain optimistic about the medium-term future. The survey suggests that executives anticipate productivity improvements of around 1.4 per cent over the next three years, along with an increase in output of approximately 0.8 per cent. However, when it comes to employment, there is a telling divergence: companies anticipate a reduction in jobs of around 0.7%, while employees themselves believe that AI will actually create a net increase of approximately 0.5% in available positions. Two groups, same technology, very different expectations.
The gap between what AI promises and what it currently delivers is not necessarily a reason to abandon it, but it is a reason to approach it with greater clarity. Real productivity gains, when they occur, will likely require more than just access to the tools. They will require careful integration, adequate training, and a willingness to honestly evaluate what works and what does not.
So far, evidence suggests that most companies have not yet reached that point.
Source: El Chapuzas Informático, February 2026. Original reporting by el-brujo. Data sourced from a National Bureau of Economic Research survey of over 6,000 financial directors and executives across Europe and the United States.
