Spain's AI jobs squeeze arrives as a hiring freeze: 40% of firms plan smaller payrolls, mostly by not hiring

The 2026 Barómetro de Talento Digital, produced by DIGITALENT Group's Observatorio del Talento Digital and presented by Nacho de Pinedo at the Digital Talent Summit at ISDI Madrid, puts a number on something Spanish employers have mostly been saying off the record: 39.6% of companies anticipate reducing headcount because of AI's expansion, as reported by El Confidencial Digital. Two adjacent figures give it shape. 77.8% report medium, high or very high AI adoption — and 79.2% admit their own professionals are poorly prepared to use it.

The most consequential detail is how the reduction arrives. Of the firms expecting a smaller payroll, 35.4% say it will come through less hiring, against 17.1% pointing to direct cuts, per El Obrero's analysis. That two-to-one ratio changes the politics of the whole thing. Layoffs generate filings, union responses, press coverage and a data trail. A hiring freeze generates nothing: vacancies that are never posted, internships quietly cancelled, a headcount plan revised down in a spreadsheet. The labour market tightens months before any indicator registers it, and the people affected never learn they were affected.

On who is exposed, the barometer confirms the pattern we have tracked sector by sector. Administrative staff are named by 50.7% of companies as the most threatened profile, followed distantly by technical roles (10.1%) and middle managers (6.8%). Senior management registers 0.5% — the layer that decides on automation concentrates the risk at the base. This is not a Spanish peculiarity. The US Bureau of Labor Statistics, in projections published on August 27, 2026, expects administrative work to be the category with the steepest percentage decline, at -4% through 2035, driven by "the continued integration of automation tools, including AI-based ones," as reported by USA Today. In absolute terms: roughly 200,600 cashier positions, 156,200 office clerks and 114,100 secretaries and administrative assistants lost between 2025 and 2035. Two labour markets, different institutions, same first casualty.

The structural damage, though, is what the study calls the junior squeeze: 51.7% of companies foresee fewer opportunities for junior profiles and another 31.4% expect significant changes to their tasks; 43.5% identify entry-level roles as the most exposed to AI, against 14.5% who point to older professionals. Here is the trap, and it is not obvious from a quarterly P&L. The tasks AI absorbs first — preparing reports, cross-checking data, handling tickets — are precisely the tasks through which people have always learned the job. Automate the apprenticeship and you stop producing your own future seniors. De Pinedo frames it this way, per El Obrero: "The challenge for companies is no longer whether to adopt AI, because in many cases it is already being used. The question is how to govern it, how it changes hiring and how professionals learn." The pipeline does not break from the top. It breaks from the bottom, and the shortage shows up five years later, when there is nobody left to poach.

The governance numbers complete the picture. 87% of professionals say they would use a personal AI tool if their employer does not provide a corporate alternative, while only 29.5% of companies have a formalised and communicated framework and 34.8% operate on informal guidance or none at all. That gap is shadow AI, and it is already costing companies leaked code, customer data and intellectual property. The report's own conclusion is the right one: banning without providing an alternative does not stop usage, it just moves it to an unlogged laptop. Microsoft, Google and LinkedIn debated exactly this at the September 22 Digital Talent Summit, per ExtraDigital, with Paco Salcedo (Microsoft Spain), Lino Cattarruzzi (Google Spain and Portugal) and Ángel Sáenz de Cenzano (LinkedIn).

Our reading: start with the caveat, because it matters. This is a survey of intentions, not a measurement of outcomes. Companies are forecasting cuts against a capability that, by their own admission in the same survey, four in five of their staff cannot yet use well. Some of that 39.6% is genuine productivity; some of it is a CFO reading headlines and pre-emptively pruning a budget line. Those two things will look identical for about eighteen months and then diverge sharply — and the firms that cut on the headline rather than on measured output will be the ones hiring back at a premium.

The deeper point is that the transition is being managed badly, not that it is unnecessary. Compressing three administrative roles into one automated flow with a human supervisor is legitimate and largely inevitable; we have said so for every sector we have covered. But the 51.7% squeezing juniors are making a different decision — trading a small, visible saving now for a talent shortage they will own entirely in five years. The companies worth watching are the ones deliberately preserving apprenticeship: keeping entry-level hires and redefining the role around supervising, auditing and directing AI rather than doing the task by hand. Long term, we still think this technology pushes toward abundance and toward work people actually choose. Getting there requires that somebody keeps training the humans in the meantime, and right now the barometer says most Spanish companies have not thought about who.