.svg)
This is the second in a two-part series on AI for doctors' offices in Switzerland and Liechtenstein. The first article covered what medical professionals are already doing with AI today: practice analytics in Excel, pseudonymized admin workflows, digital twins for patient journeys, and voice-to-report tools. If you have not read it, we recommend starting there. This article covers what comes next: a five-week course designed specifically for doctors' offices in the region, why it exists, who built it, and what participants walk away with. Here's the link to register for the course
In the first article, we described a capacity trap: practices that are too overloaded to adopt the tools that would reduce the overload. We also described a trust gap: the tools work, but nobody has shown doctors how to use them within the data protection standards they operate under.
These two gaps define the problem. The bottleneck is guidance.
Dr. Alberto Tozzi, who joined one of our sessions from Bambino Gesù Children's Hospital, made an observation about medical education and AI that stayed with us: clinician-to-clinician communication is far more effective than engineer-to-doctor communication. You start from the clinical case, not the technology. If you start from the engineering concepts, the message does not land.
Most AI courses do the opposite. They teach tools, frameworks, and concepts, then hope participants figure out the context. For a doctor's office in eastern Switzerland operating under Swiss and Liechtenstein data protection standards, with specific practice management software, specific billing workflows, and 2.5 hours of daily admin eating into patient time, generic AI training is not useful. It needs to start from the specific problems these practices face. That is what this course was built to do.
The brains behind the course: MPA Berufs- und Handelsschule AG and werchota.ai
MPA is a medical education institute in eastern Switzerland. It has been operating for over 30 years and holds EDUQUA certification, the Swiss quality standard for professional and continuing education. The school trains medical practice assistants and offers further education programs for healthcare professionals. MPA works directly with over 300 doctors' offices in eastern Switzerland and Liechtenstein. These are the practices that send their staff for training, that the school's leadership speaks to regularly, and whose daily constraints shaped the course syllabus.
The school is led by Gernot Boiger, who took it over with his wife after spending 12 years as a professor of computational physics at the Zurich University of Applied Sciences (ZHAW), where he remains active in a part-time capacity. His background in computational modelling and engineering shaped MPA's approach to AI in ways you would not expect from an education institute.
Under his leadership, MPA has begun developing its own AI-powered tools. Pathfinder is an adaptive learning platform that tracks individual student progress and allows teachers to set up personalized learning paths. It is funded by the Gebert Rüf Foundation and already in pilot use with students. Smart Admin is an AI-powered system to modernize the school's own administrative processes. Both were built in-house. MPA builds and uses AI in its own operations.
werchota.ai is an AI consultancy working with companies across the DACH region on practical AI adoption. The company brings fluency with the current generation of AI tools: Microsoft Copilot, Claude, voice-first workflows, and rapid prototyping through vibe coding. The team uses AI daily for content production, client work, internal administration, and software development. When co-founder Malcolm Werchota demonstrates Copilot in Excel to a room of medical professionals, he is showing a tool he uses himself.
The partnership exists because neither organization could build this course alone. werchota.ai has the tool fluency and the knowledge of what is new, what is compliant, and what is practical. MPA has the relationships, the regional trust, and the understanding of what a doctor's office in this part of Switzerland actually needs. The school has been embedded in this community for three decades. That trust is the delivery mechanism.
Gernot and Malcolm have known each other for 20 years, since their student days in Leoben, Austria. Their careers diverged across continents and industries, but when they found themselves 30 kilometers apart in eastern Switzerland, the course designed itself. Malcolm asked one question: where could AI benefit your existing customers the most? The answer was immediate: doctors' offices.
A sneak-peak into the syllabus of the course
The course is called *Sichere KI-Anwendung in der Arztpraxis* (Safe AI Application in the Doctor's Office). It runs for five weeks.
Malcolm Werchota leads the theory and tool sessions. These cover the current AI landscape, which tools are suitable for medical practice use, how to evaluate them for compliance, and how to build workflows that respect data protection requirements. Gernot Boiger leads the practice sessions on-site at MPA's seminar space in eastern Switzerland. These are hands-on: participants work with real practice scenarios, build prototypes, and apply what they learned in the theory sessions to their own context.
The format is hybrid. Theory sessions are delivered online, projected into the physical seminar space so on-site participants see Malcolm alongside them. Practice sessions happen in person, with breakout rooms for small-group work.
AI inside the tools you already use: How to use Copilot in Excel, Outlook, and the broader Microsoft stack for practice analytics, scheduling insights, billing analysis, and inventory management. The first article in this series showed what a single spoken prompt in Excel can produce in eight minutes. The course teaches participants how to do this with their own data.
Voice-first workflows: Doctors dictate constantly. Voice is their natural interface. The course covers how to use voice-to-AI workflows for documentation, reporting, and administrative tasks, building on skills doctors already have.
Pseudonymization techniques: How to encode patient data before it reaches an AI model, so that identifying information (names, birth dates, AHV numbers) never leaves the practice's control. The first article described a practice in our network already doing this successfully. The course teaches the method.
Compliance evaluation: How to assess whether a specific AI tool meets Swiss and Liechtenstein data protection requirements. What questions to ask. What infrastructure to look for (private tenants, data residency, enterprise agreements). How to distinguish marketing claims from actual compliance.
Building a working application: Each participant builds a working prototype for their own practice during the course, addressing a specific problem in their daily operations. Gernot guides this process on-site, drawing on his engineering background and his understanding of practice workflows from MPA's 30-year relationship with the local medical community.
Why "Sichere" is in the title
The word "Sichere" (safe) is deliberate. In eastern Switzerland and Liechtenstein, data protection is a cultural expectation that patients, staff, and the broader community hold practices to. Liechtenstein implements the full GDPR through the EEA, with penalties up to many millions. Switzerland's own data protection culture often exceeds what its laws require on paper.
The course was designed around these standards, not in spite of them. Every tool, technique, and workflow taught in the course has been evaluated against the data protection requirements of this specific region. Pseudonymization is built into how participants learn to work with AI from the first session.
The reasoning is straightforward: if you can teach AI adoption under the strictest data protection culture in Europe, using tools doctors already have, guided by people they already trust, the approach works anywhere.
Who this course is for
The course is designed for doctors, medical practice assistants, and practice managers in the DACH region who:
- Know that AI could reduce their administrative load but have not found a safe way to start
- Need to understand what is GDPR-compliant and what is not, in practical terms
- Want to build something for their own practice, not just learn concepts
- Work within Swiss or Liechtenstein data protection standards and take those standards seriously
A practical program for people who want to apply AI in their specific working environment, with the specific constraints they operate under, and walk away with something that works.