McKinsey reveals that 60% of corporate jobs are already experiencing a direct transformation due to artificial intelligence in 2026, while 45% of work activities could be automated. However, less than 5% of occupations would disappear completely.
The impact of artificial intelligence on corporate employment is already a quantifiable reality. According to the latest report from McKinsey, 60% of jobs in companies are facing a direct transformation due to AI this year. Additionally, up to 45% of work activities could be automated, although less than 5% of occupations would disappear completely.
For founders and leaders of startups, this data does not represent an existential threat, but rather an opportunity to restructure teams before the competition does. The difference between companies that survive and those that scale in 2026 no longer lies in who adopts AI first, but in who better prepares their talent for the new roles that automation creates.
The global figures are compelling. BCG estimates that in the next two to three years, between 50% and 55% of jobs in the United States will be reconfigured by AI. Within a five-year horizon, between 10% and 15% of current positions could be eliminated.
PwC, in its AI Jobs Barometer 2026, analysed over a billion job postings across six continents. The pattern is revealing: companies most exposed to AI are increasing both salaries and headcounts more rapidly. Automation is not net destroying jobs but displacing them towards roles that require human oversight, process management, and hybrid skills.
Goldman Sachs projects that widespread adoption of AI would take around ten years, displacing 6%-7% of workers during that transition. The World Economic Forum goes further: by 2030, 170 million new jobs are projected alongside 92 million displaced, resulting in a net creation of 78 million positions.
In Spain, data from Funcas (2026) reveals that between 1.7 and 2.3 million jobs could be affected by AI between 2025 and 2035, with a potential creation of 1.61 million new jobs linked to artificial intelligence technologies.
The roles most pressured by automation are those with routine, repetitive, and standardisable tasks. According to multiple studies, the most exposed include administrative and office technical roles, junior data analysts, programmers engaged in repetitive tasks, writers of standardised content, traditional sales and customer service representatives, information processors, and routine operators in logistics.
A study released by Microsoft identified 40 professions most vulnerable to automation by AI by 2026. The key is not the job title, but how many of its tasks can be almost entirely performed by automated systems. When AI can perform the majority of a role's tasks, the presence of that position within a company drops by nearly 14%, according to research from MIT Sloan.
In Mexico, a critical data point emerges from recent research: 51% of workers believe that the incorporation of AI in the workplace will primarily benefit companies, not them. This perception gap creates resistance to change just when collaboration between human teams and automated systems is most needed.
The automation of processes is shaping up to be the main application of AI in Mexican companies during 2026. However, the region presents a unique opportunity: Latin American markets have fewer legacy systems than the United States or Europe, allowing for the implementation of AI workflows from scratch rather than migrating outdated systems.
In sectors such as manufacturing, logistics, warehousing, agribusiness, and mining, the convergence of AI, process automation, physical robotics, and data is already generating measurable efficiencies. Startups that understand this convergence ahead of established companies will have a competitive advantage.
For founders, the window to prepare the team is open now, but it will close once competitors have completed their transformation. Concrete actions include identifying automatable tasks within the organization and redefining roles towards supervision and management of AI-assisted processes.
You may also like
