AI and employment in Switzerland: first job losses – a clear warning sign for SMEs

Artificial intelligence is no longer just a tool being trialled in a corner by marketing or IT teams. In Switzerland, it is beginning to play a role in reorganisation decisions. Le Nouvelliste, citing Keystone-ATS, reports that the first job cuts attributed to AI have been announced in the country, with companies such as DocMorris and Lastminute.com highlighting the expected efficiency gains.

For SME leaders, this is an important signal, but it should not be interpreted as a mechanical wave of redundancies. The available data instead points to a more subtle transformation: AI is changing tasks, the types of roles sought and the organisation of work before causing massive upheaval in the workforce. It is precisely this grey area that deserves the attention of employers, HR managers and payroll service providers: the balance between promised productivity gains, social risks, data protection and the need for training.

Highly visible, yet still targeted, job cuts

The most concrete example mentioned in the Swiss press is that of DocMorris. According to Le Nouvelliste, the Swiss online pharmacy announced on 25 June that it was cutting around 100 jobs across the group as part of a strategic reorientation giving greater prominence to artificial intelligence. The article, published on 2 July 2026 and sourced from Keystone-ATS, also cites Lastminute.com among the companies relying on AI to improve their efficiency.

These announcements are striking because they directly link a technology to job losses. They mark a turning point in corporate economic communication: AI is becoming a productivity argument presented to markets, investors and employees. Le Nouvelliste also notes that some experts sometimes see this as a way of reassuring investors. In other words, not all job cuts announced in the name of AI necessarily mean that machines are immediately replacing specific employees, task by task.

For an SME, this distinction is crucial. The introduction of an AI tool can reduce the time spent on drafting, sorting information, translating, responding to simple enquiries or preparing documents. But turning this time saving into job cuts requires a comprehensive organisational analysis: which tasks are actually being phased out, which still require human validation, what new checks are being introduced, and what level of quality do customers expect?

In SMEs, adoption is outpacing redundancy plans

Figures available from Swiss SMEs show rapid adoption. According to AXA’s study on the SME labour market, the proportion of companies integrating AI into their processes rose from 22 per cent to 34 per cent between 2024 and 2025. At the same time, the proportion of SMEs that have never used AI fell from 45 per cent to 29 per cent.

The shift in perception is just as marked. In 2025, 60 per cent of SMEs view AI as an opportunity, compared with 35 per cent in 2024. The proportion of those who see it as a threat has fallen to 8 per cent, down from 20 per cent the previous year. AI is therefore rapidly becoming commonplace in the management of small and medium-sized enterprises, no longer as a futuristic concept, but as an everyday business tool.

However, this adoption has not yet led to a massive reduction in staff numbers within SMEs. According to AXA, 2 per cent of SMEs have reduced their workforce in 2025 thanks to productivity gains linked to AI, whilst 10 per cent have created new roles. The key issue is therefore not just the number of jobs, but the nature of the skills required. A company can maintain its workforce whilst fundamentally changing the expectations placed on staff: the ability to use an AI assistant, to monitor its results, to formulate clear instructions, to spot a plausible but incorrect answer, or to integrate the tool into a business process.

In practical terms, AI is already widely used in office tasks. The AXA study specifically cites translation, used by 52 per cent of SMEs that use AI, and correspondence, mentioned by 47 per cent. For a company operating in several national languages or dealing with an international client base, these applications can save time. But they also shift the focus of the work: it is no longer simply a matter of producing a text, but of checking the tone, the meaning, the data used and how well it suits the client’s situation.

Administration, management, support: office roles on the front line

Exposure to AI does not automatically mean the disappearance of a job. Rather, it indicates that a significant proportion of a profession’s tasks can be assisted, accelerated or partially automated. According to a study by Kuble and Employés Suisse cited by Le Temps, 28 per cent of jobs in Switzerland are highly exposed to AI, particularly in the fields of administration and management.

For SMEs, this observation is particularly relevant. Administrative, HR, accounting, sales and customer support roles often involve handling information: reading, filing, summarising, comparing, drafting, following up and documenting. These are precisely the activities in which generative AI tools can be useful. In an accounts department, for example, AI can help draft an email, summarise an internal memo or structure a checklist. In an HR team, it can assist with drafting a job advertisement or preparing a training document. However, responsibility for decision-making, compliance and human relations remains with the organisation.

The risk for management would be to confuse the automation of a task with the automation of a role. An employee is not generally limited to the actions visible within a software programme. They know the clients, interpret exceptions, recognise when a request falls outside the usual scope, pass on informal information and sometimes protect the company from costly mistakes. AI can reduce certain repetitive tasks, but it can also create new requirements: configuration, proofreading, documentation, quality control, cybersecurity, access management and staff training.

The real HR challenge: reorganising before cutting back

When an SME introduces AI, the first question should not be: how many jobs can we cut? A more robust approach is to map out the tasks. Which activities are repetitive? Which ones require business expertise? Where would an error have a significant impact on a customer, an employee, a declaration, a payslip or a business decision? This analysis helps to distinguish between easy wins and areas where caution is required.

In practice, AI can take on some of the preparatory work: producing a first draft, suggesting a structure, summarising a conversation, rephrasing a text, or identifying apparent inconsistencies. But it does not replace professional judgement. For a Swiss SME, value often lies in proximity to clients, knowledge of the local community and the ability to handle specific cases. If AI standardises responses too much, it can also undermine what sets the company apart commercially.

Internal dialogue therefore becomes crucial. Introducing AI without explaining the objectives fuels anxiety. Introducing it with clear rules, on the other hand, helps to engage teams: which tasks do we want to simplify, which checks should be retained, which skills should be developed, and which uses should be prohibited? Staff who are familiar with the processes are often best placed to identify tasks that can be automated without compromising quality. Involving them at an early stage can reduce resistance and prevent decisions being made solely on the basis of a supplier’s promises.

In the event of a reorganisation affecting jobs, caution is essential. Decisions must be assessed in the light of employment contracts, the employer’s obligations, internal practices and the legal framework applicable to the specific situation. AI does not replace the need for a rigorous HR, legal and financial analysis. Before announcing a job cut “thanks to AI”, an SME should be able to demonstrate what is actually changing in the work and how the remaining tasks will be carried out.

Data, usage rules and training: the less spectacular challenge

The issue is not limited to employment. The use of AI also raises questions of governance. According to AXA, only one in three SMEs has put in place a clear data protection policy relating to AI. This is a sensitive issue: AI tools are often used spontaneously by staff, sometimes before the company has defined what data may be entered into an external system.

A simple rule should guide SMEs: do not treat AI as a personal gadget, but as a business tool. This involves specifying permitted uses, the data that must not be entered, the necessary approvals before sending information to a client, and responsibilities in the event of an error. Salary information, customer data, contractual details and strategic documents warrant particular attention. Even without seeking detailed legal advice, a company would be well advised to have its practices reviewed on a case-by-case basis by competent specialists.

Furthermore, the Swiss regulatory framework is still under development. The research report notes that Switzerland does not yet have specific legislation governing AI. Discussions are, however, underway, particularly at cantonal level: according to PwC, Geneva has commissioned an analysis of the impact of AI on its economy. This lack of dedicated legislation does not mean there are no responsibilities. Rather, it requires companies to proceed methodically, combining caution, documentation and sound operational judgement.

For business leaders, the best immediate investment is not always the most advanced software. It is often training: teaching teams how to use AI without blindly delegating decision-making to it, how to verify results, how to protect sensitive information and how to recognise the tool’s limitations. AI can save time, but the benefits only become sustainable if the organisation knows where to apply human oversight.

The job cuts announced in Switzerland show that AI is now a factor in companies’ strategic decision-making. But the figures from SMEs suggest a less black-and-white interpretation: AI eliminates certain needs, creates others and, above all, redefines the skills required. For Swiss employers, the challenge is not to choose between enthusiasm and fear, but to steer this transition before it is imposed upon them — with rules, reliable internal data and a clear vision of the work the company still wishes to entrust to humans.