Artificial intelligence and the patient journey in Switzerland: 2025 insights and opportunities for your practice

In just a few years, artificial intelligence (AI) has gone from being a tech curiosity to a tool used by the general public. But how does it really fit into healthcare? In this article, we break down the findings from our 2025 study on how Swiss patients are using AI, the hurdles that still exist, and most importantly, the concrete opportunities it offers your medical practice.

1. Are Swiss patients really using AI?

Since ChatGPT launched at the end of 2022, artificial intelligence has become increasingly mainstream in Switzerland. A Comparis survey published in March 2025 shows that two-thirds of adults have already interacted with a generative chatbot like ChatGPT or Gemini, and 98% of the population is familiar with these tools according to a study by the University of Zurich conducted in late 2024.

However, when it comes to health, the Swiss population seems hesitant: our study, carried out with Farner and MIS Trend, shows that 74% do not use AI in their healthcare journey.

AI use in healthcare remains limited

AI tools are especially underused by people over 55 and those with a EFZ/EBA diploma. Among the quarter of the population who have tried it, the main uses of AI in healthcare are:

• Asking a medical question to a chatbot (16%)
• Comparing or confirming a diagnosis (11%)
• Interpreting test results (11%)
• Checking a prescription (7%)

In other words, even curious patients mainly use AI as a complement or second opinion — not as a replacement for their doctor.

Trust in AI in medicine is still fragile

Only 35% of respondents said they trust AI even “a little” to diagnose or recommend treatment. Just 2% said they are “completely confident.” Trust decreases with age and is generally lower among women.

People already using digital health tools (like online appointment booking or the electronic patient record) tend to be twice as confident in AI.

The main reasons for this caution are fear of serious errors, lack of transparency around medical sources, and concerns over personal data privacy. Furthermore, no AI solutions are currently officially recognized in Switzerland for medical diagnostics.

3. Is AI a threat to healthcare professionals?

Not at all. As the results show, most people in Switzerland approach the use of AI in healthcare with caution and still prioritize human diagnosis. In fact, our study found that three out of four Swiss people don’t believe AI will replace their doctor. This belief is even stronger in German-speaking Switzerland (80%) than in Romandy (63%) and Ticino (64%). As we’ve seen, trust in AI for diagnosis remains very low.

2. Can AI help reduce healthcare costs in Switzerland?

What patients think

When asked, “Can AI help reduce healthcare costs?”, 61% said “yes” — but only 11% are fully convinced, mostly younger adults (18–34 years old).

Where people see potential savings

When asked how AI could lower costs, the most common answers were:

1. Digitalization and secure sharing of health data: nearly half believe that automating data flows would help avoid duplicate tests and save time.
2. Checking or confirming a diagnosis: 53% think AI could help reduce errors and therefore unnecessary treatments.
3. Faster diagnosis through image or signal analysis, mentioned by 47%.
4. Telemedicine and emergency triage: 40% believe faster referrals to the right level of care could reduce unnecessary consultations and hospitalizations.

AI as a valuable ally for Swiss practices

AI should not be seen as a threat or competition for healthcare professionals. Swiss patients still trust human diagnosis far more than AI — and that’s not likely to change any time soon.

That said, AI can still play a role in your practice! It can be a real asset when it comes to handling complex administrative tasks. As is often the case, the key lies in striking the right balance between technological support and the human touch.

Study methodology

This survey was conducted online by OneDoc, Farner, and MIS Trend in March 2025, among a representative sample of 1,124 Swiss residents aged 18 and over (396 in French-speaking Switzerland, 434 in German-speaking Switzerland, and 294 in Ticino). The maximum margin of error is ±2.9% nationwide. The data were weighted to ensure regional and linguistic representativeness.

Discover all the findings from our study in our press release (DE)!