In Switzerland, workplace upheavals linked to the rise of Artificial Intelligence (AI) are particularly pronounced. The reason for this is a focus on highly knowledge-based fields and concentration of company headquarters. This also applies to countries such as the UK and the Netherlands.
According to a recent analysis, 18 percent of the tasks incumbent upon employees in Swiss businesses are in areas that can be automated by harnessing the power of Artificial Intelligence (AI). Alongside the UK, the Netherlands, Luxembourg and the Republic of Ireland, this puts Switzerland in a group of countries that is particularly exposed to the potential ramifications of the AI revolution.
This insight comes from a study conducted by Coface in partnership with the French research institute OEM (L'Observatoire des emplois menacés et émergents). In this study, we analyzed the potential for AI-driven automation across the labor markets of nearly 30 advanced economies and identified the occupations most likely to be affected by these developments.
Why knowledge workers face greater AI disruption than production jobs
The study assumes that modern AI applications are causing a significant shift in the boundaries for automation, where AI is increasingly capable of handling cognitive, complex and skilled tasks. As a result, experts defined tasks and basic actions for 923 professions and weighted them according to their importance and frequency.
The findings reveal that AI is transforming the employment structure on a profound level: More than a quarter of work tasks in management and administration, creative professions, law and finance, as well as engineering and IT could reportedly be automated. Individual services, technical, skilled trades and industrial production jobs fall below a 10 percent automation threshold. On this basis, a geographical picture was created to distinguish between different phases of AI development, taking into account not only sectoral differences but also local business structures.
How Switzerland ranks globally: 4th most exposed out of 30 countries
The high-exposure country cluster, of which Switzerland is also part, is characterized by numerous company headquarters. Here, the proportion of exposed corporate functions and professional groups is accordingly high, particularly in relation to finance, IT and law. The proportion of tasks that can be automated using AI models stands at 21 percent in Luxembourg, which leads this group, while rates of nearly 20 percent and around 19 percent are recorded for the UK and the Netherlands respectively. After Switzerland at 18 percent, Ireland follows closely behind at 17 percent.
The study also identifies a Nordic cluster of welfare states, which is impacted to a slightly lesser degree, comprising Norway and Iceland, for example, but also tends to include states such as Estonia. A Central European cluster around Austria, Slovenia and Czechia shows similarities with Germany in terms of a strong focus on engineering, but also a greater emphasis on production. However, the southern cluster is even less exposed to AI automation, whereby trade, individual services, construction and transport play a larger role in places such as Portugal, Italy, Spain and Romania. For example, only 12 percent of work content in Turkey is potentially subject to automation.
AI won't destroy jobs – but it will fundamentally reshape them
Exposure does not, however, automatically mean job losses.
as Markus Kuger, Coface economist responsible for the DACH (Germany, Austria, Switzerland) region, explains in a statement.
The Coface approach measures the technical exposure of tasks to automation and not the actual net effects on employment. Nevertheless, he does issue some words of warning: “Demand effects could have an impact, new jobs may emerge or obstacles may come into play”.
According to Kuger, in Switzerland the first step is likely to involve reallocating tasks within existing roles. For example, junior or back-office positions may involve fewer routine tasks, while productivity pressure could ramp up. Moreover, there will be an increasing emphasis on judgment, expertise, governance and the monitoring of AI systems. It is only further down the line that jobs in knowledge-based fields and employment structures could potentially undergo long-term transformations.
Who really benefits? Infrastructure power sits with a handful of foreign players
For the experts, the key question is who actually benefits from the value generated.
Swiss companies can increase productivity, but the infrastructure level – chips, data centers, cloud services and foundation models – is controlled by a small number of predominantly foreign providers.
Kruger explains.
In this way, added value is lost, while new geopolitical dependencies and risks in the areas of cybersecurity and regulation come to the fore.
In a scenario crucial to the study, where agentic AI workflows become widespread, there would be a tangible shift in value creation from labor to capital over the long term. When we consider GDP per capita in relation to the potential degree of task automation, Luxembourg is clearly ahead in first place, with Switzerland and Ireland following. This development could therefore have far-reaching consequences if tax systems are weighted more towards taxing labor. In such cases, it must be expected that revenues from social security contributions or income taxes will decrease, while the bill for unemployment insurance and further training expenditure will grow.
At the same time, a debate focused on the value of education and formal qualifications is taking place. In this context, it is already evident that degrees in the traditional academic sense are starting to lose value.
