AI for practice management in Switzerland: where it helps, what the rules say and how to start
By AllesHealth GmbH ·
Most talk about AI in medicine is about diagnosis or AI scribes. But if you run a practice, the questions are often more ordinary: who answers the messages that arrived overnight, who sends the reminders, who drafts the referral letter, who checks whether a patient might fit a study? This article looks at where AI can realistically help with practice management in Switzerland, what Swiss rules apply as of November 2026, and how to adopt it without losing control.
If you are mainly interested in AI that writes consultation notes, read our separate article on AI scribes and AI documentation. Here the focus is on running the practice as a whole.
Why practice management is where the time goes
Administration is a real burden for Swiss doctors. In the survey that gfs.bern conducted for the FMH in 2025, with 1,532 doctors taking part, doctors in acute-care hospitals reported spending an average of 114 minutes a day on work on the patient record, and doctors in psychiatry 121 minutes a day on medical documentation [1]. Doctors working in private practice reported spending an average of 54 minutes a day on administrative tasks arising from the requirements of authorities and insurers [1].
Health professionals are open to digital help if it removes work. The Swiss eHealth Barometer 2026 surveyed 1,187 health professionals and other healthcare stakeholders. For the planned Swiss health data space (SwissHDS), they expect more efficient information flows, better availability of patient data and relief from administrative work; their willingness to take part is high but depends on clear added value without extra work [2].
That is the right test for any AI tool in a practice: does it take a task off someone's desk, or does it just move it?
Where AI realistically helps in a Swiss practice
The tasks below are repetitive, rule-based and high in volume. That makes them better candidates for AI support than complex clinical judgement.
Scheduling and online booking
AI can help patients find a suitable slot, suggest appointment types based on the reason for the visit and fill gaps from cancellations. The practice still defines the rules: which appointment types patients may book themselves, how long each one takes, and which requests always go to a person.
Appointment and medication reminders
Reminders are one of the best-studied low-tech tools in outpatient care. A Cochrane review found low- to moderate-quality evidence that text message reminders increase attendance at healthcare appointments compared with no reminders [3]. AI adds value by sending reminders automatically, at the right time and in the patient's language, and by making it easy to cancel or rebook.
Inbox and message triage
Messages from patients, labs and other providers arrive in one stream. AI can sort them by topic (prescription request, appointment, result question, administrative), flag messages that look urgent for immediate human review and suggest a draft reply for routine questions. A person decides what is sent.
Documentation drafts
Beyond the consultation note, practices write referral letters, reports for insurers and summaries for patients. AI can prepare a first draft from data already in the record. The doctor checks it, corrects it and signs it. See how to stop charting after hours for workflow changes that matter at least as much as the tool.
Coding support
AI can suggest structured codes, for example SNOMED CT concepts for diagnoses, based on what has been documented. This makes the record easier to search and reuse. The suggestion is a proposal; the clinician confirms it.
Finding eligible clinical trials
Matching patients with clinical trials usually requires checking long eligibility criteria by hand. AI can compare structured data such as biomarkers with trial criteria and suggest patients who may be eligible. The doctor decides whether to raise the option with the patient.
Risks and rules: what applies in Switzerland
Data protection applies to AI today
The Federal Data Protection and Information Commissioner (FDPIC) states that the revised Federal Act on Data Protection (FADP), in force since 1 September 2023, is directly applicable to AI-supported data processing [4]. Manufacturers, providers and users of AI systems must make the purpose, functioning and data sources of AI-based processing transparent [4]. People who interact with a chatbot have the right to know that they are dealing with a machine, and for processing with high risks the law requires a data protection impact assessment [4].
For a practice this means: tell patients when they are communicating with an automated system, document which AI functions you use and why, and assess the risks before you introduce a tool that processes health data.
Where the data is hosted matters
Under Swiss law, personal data may in principle be disclosed abroad freely only to countries that the Federal Council lists as having an adequate level of data protection [5]. For other countries, suitable safeguards such as contractual data protection clauses or standard data protection clauses are needed, unless one of the exceptions in the law applies [5]. Ask every vendor where data is stored, where the AI models run and which subcontractors are involved.
Is an AI tool a medical device?
According to Swissmedic, the intended purpose defined by the manufacturer decides whether software falls under the medical device regulations [6]. Software for hospital resource planning, reimbursement or managing doctors' visits, and electronic records that simply replace paper, are listed as examples of software that is not a medical device [6]. Software can qualify as a medical device when it is intended for a medical purpose for an individual patient and processes medical data beyond storage, archiving, communication or simple search [6]. If software consists of several modules, the manufacturer is responsible for qualifying and classifying them, either as a whole or module by module [6].
So a scheduling assistant is usually not a medical device, while a tool that suggests diagnoses or treatment may be. Ask the vendor how each AI function is qualified.
No general AI law yet
As of November 2026, Switzerland has no overarching law specifically on AI [7]. In February 2025 the Federal Council decided to ratify the Council of Europe's AI Convention and to adapt Swiss law where needed, as far as possible sector by sector [7]. Check for updates before major investments.
Human oversight is not optional
AI drafts can be wrong, incomplete or written with false confidence. The doctor remains responsible for every entry in the record, every message with clinical content and every decision about a patient. Build review into the workflow so it happens every time, not only when someone has time.
A practical adoption checklist
- Start with one bottleneck. Measure where time goes for two weeks (phone, inbox, letters, reminders) and choose the task with the highest volume and lowest clinical risk.
- Define what AI may and may not do. For example: it may draft replies to appointment requests; it may not send any message with clinical advice without review.
- Check data protection before go-live. Hosting location, subcontractors, transparency for patients and, where risks are high, a data protection impact assessment [4][5].
- Ask about medical device status for any function that goes beyond administration [6].
- Secure the basics. The FMH has published eleven IT security recommendations for practices with up to twelve doctors, including backups and two-factor authentication [8]. AI does not replace them.
- Train the whole team. MPAs and nurses often use the AI functions most. They need to know how to check outputs and when to escalate.
- Review after three months. Has the task really become lighter? Are errors caught? Keep what works and drop what does not.
How to choose the best practice management software with AI in 2026
There is no single "best" product for every practice. Use criteria instead:
- AI where the work is: booking, reminders, message handling, documentation drafts, coding support, not just a chatbot on the website.
- Clinician in control: every AI output is a draft or suggestion that a person reviews, and it is clear what was generated automatically.
- Integrated, not bolted on: AI works inside the record, the calendar and the patient app, so data is not copied between systems.
- Swiss hosting and transparency: clear statements on where data is stored and processed, and on subcontractors.
- Standards: structured data (for example HL7 FHIR, SNOMED CT) so information can be reused and exchanged.
- Swiss billing: support for Swiss tariffs and electronic billing to health insurers.
- Low barrier to try: a trial period in your own practice says more than any demo.
For a broader view, see the administrative burden on doctors and our practical guide to reducing paperwork in a medical practice.
How AllesHealth Connect helps
AllesHealth Connect is practice software for GP and multi-doctor outpatient practices, designed in Switzerland. It combines the electronic health record, appointment management, e-prescriptions, video consultations, and billing and invoicing with support for Swiss tariffs and electronic billing to health insurers. In the Premium plan, AI agents take on routine work. AI drafts and suggests; the clinician reviews every result and stays responsible.
For doctors
- AI-assisted EHR updates, including voice dictation, so documentation can be drafted while the consultation is fresh and then checked.
- AI clinical-trial matching based on biomarker data, so potentially eligible patients are flagged for your review.
- Structured data built on HL7 FHIR with SNOMED CT coding support.
- Lab results arrive electronically where the lab is integrated.
For nurses and MPAs
- AI agents for booking and reminders take on routine scheduling work.
- Appointment management keeps the team's calendar in one place.
- Messages with patients run through one channel in Patient Companion instead of scattered calls and emails.
For patients
- Patient Companion, the app in your practice's own branding for iPhone and Android: online booking, video consultations, medication and dosage reminders, messages with the team, and a library of prescriptions, lab reports and imaging.
- Family members or carers can be managed in one account.
Data of Swiss practices is hosted in Switzerland (Zurich). Plans are available for practices of different sizes; see the pricing.
Try AllesHealth Connect free for 60 days – no setup needed, no credit card: start free or book a demo.
#AIinHealthcare #PracticeManagement #ClinicManagement #DigitalHealth #GPPractice #MPA #DataProtection #Switzerland
Sources
- Befragung zum ärztlichen Arbeitsumfeld im Auftrag der FMH 2025 (Online-Cockpit mit Ergebnissen), gfs.bern (on behalf of FMH Swiss Medical Association) (2025)
- Swiss eHealth-Barometer 2026: Bericht zur Befragung der Gesundheitsfachpersonen und Akteur:innen des Gesundheitswesens, gfs.bern (2026)
- Mobile phone messaging reminders for attendance at healthcare appointments, Cochrane Database of Systematic Reviews (2013)
- Update - Current data protection legislation is directly applicable to AI, Federal Data Protection and Information Commissioner (FDPIC) (2025)
- Cross-border transfer of personal data, Federal Data Protection and Information Commissioner (FDPIC) (2024)
- Information sheet: Medical Device Software (BW630_30_007), Swissmedic (2026)
- Artificial Intelligence, Federal Office of Communications (OFCOM) (2026)
- IT-Grundschutz für Praxisärztinnen und Praxisärzte – 11 Empfehlungen (D3), FMH Swiss Medical Association (2023)