Six in ten workers now use AI. What about the other four?

There is a number that has quietly changed the conversation about artificial intelligence in the workplace: 60%. According to the most recent data for Spain, six in ten professionals now use AI tools regularly at work. Not as an experiment, not as an occasional novelty, but as part of their daily workflow.

That means four in ten do not. And that gap — which two years ago was the majority — is rapidly becoming a competitive disadvantage worth taking seriously.

The number that changed the conversation

For years, AI adoption at work was a story about large technology companies and technical roles. A 60% adoption rate in Spain — a labour market with a more traditional sector structure than the European average — signals that the inflection point has already happened.

What matters most is not the number itself, but what it says about the speed of change. In 2023, the figure was around 20-25%. In under three years, it has crossed into majority territory. Very few technologies have achieved this adoption velocity in actual workplace settings, where habits change slowly and institutional resistance is high.

The sectors with the highest penetration are the predictable ones — technology, marketing, consulting, education — but the expansion into more traditional sectors like healthcare, logistics and financial services is what makes the 60% figure genuinely significant rather than just a reflection of a technical minority.

What they are actually doing with AI

The uses that have gained the most ground are not the most sophisticated. They are the most everyday: drafting and revising emails and documents, summarising long information, preparing presentations, searching and synthesising information, translating content.

This matters because it dismantles the myth that integrating AI into work requires technical knowledge. The dominant mode of use is conversational: type what you need into a chat interface and get a result to refine. The learning curve for these basic uses is measured in hours, not weeks.

The more advanced uses — automating complete workflows, connecting tools through agents, generating functional code — are what distinguish users who save time from users who transform their role. Those require more investment, but they are not exclusive to technical profiles either.

What all users reporting the highest impact share is the same insight: they do not use AI to outsource thinking. They use it to accelerate the parts of their work that consume time without generating distinctive value. That distinction is crucial.

The four-in-ten problem

Those who do not use AI habitually are not necessarily technology-averse or resistant to change. The most common reasons that repeat are three: not knowing where to start; not having encountered a use case concrete enough to justify the effort of learning; or having tried once without a clear result and not returning to it.

The problem is that the window in which this is neutral — in which not using AI carries no cost — is closing. When 60% of professionals in your sector already operate with these tools habitually, the asymmetry in speed and capacity starts to become visible. Not all at once, but cumulatively.

In labour markets where individual productivity matters, the difference between producing in four hours what used to take eight — at the same quality — is not a minor detail. It is an advantage that compounds every week.

The gap is not technological

The biggest barrier to adoption is not technical. It is mindset and habit.

The technology is available, inexpensive — most relevant tools have free or low-cost plans — and requires no complex installation or configuration. The real obstacle is inertia: it is easier to keep doing things the way you always have than to invest the initial time needed to integrate something new into your daily workflow.

This is the same obstacle that exists with any habit change. And the solution is the same: do not try to change everything at once. Find one specific use case where the tool saves time in an obvious way, adopt it until it becomes automatic, and expand from there.

The minimum viable entry point

If you do not use AI at work and are not sure where to begin, there is a concrete recommendation that works for most people: take the task in your week that consumes the most time and that you most dislike doing, and try delegating it wholly or partly to a tool like Claude, ChatGPT or Gemini.

Not the most important work. Not the most creative. The most tedious and repetitive. That is the right entry point because the satisfaction threshold is low — anything better than the current state justifies the experiment — and the cost of a mediocre result is limited.

From that first concrete success, the rest of the learning curve becomes substantially more manageable.