Clear lines ahead

AI for Medical Practices: Relief for Phones, Mail and Admin

Monday, 8 a.m.: the phone won't stop ringing, there's a queue at reception, and the inbox is full of doctors' letters, lab results and three faxes that should really have been KIM messages. Your medical assistants are juggling appointments, repeat prescriptions, referrals and follow-up questions at once - before consultations have even started.

AI for medical practices in Germany shouldn't mean yet another portal. It should bring relief to exactly these desks. At Sailrs, we build AI solutions that answer and structure calls, pre-sort the inbox and make practice knowledge easy to find. Medical decisions always stay with you and your team: our solutions are administration and communication tools, not diagnostic software.

Where everyday practice gets stuck

In GP and specialist practices the pattern repeats: the medicine is well organised, the admin around it isn't. Typical bottlenecks we hear about again and again:

  • Phone peaks: Calls pile up in the morning and after lunch. Many are routine - repeat prescription, referral, rescheduling, "Are my results in?" - yet they still block the line for urgent matters.
  • Media breaks in the inbox: Doctors' letters, lab values and hospital discharge letters arrive via KIM, fax, post or patient portal. Someone has to read, assign and transfer them into the practice management system (PVS).
  • Knowledge in heads and binders: Quality manual, hygiene plan, billing notes for EBM and GOÄ, work instructions for new colleagues - it all exists, but it's rarely at hand the second someone asks.
  • Staff shortages: When one assistant is off sick, the balance tips immediately. Every routine task that no longer needs manual handling creates a buffer.

This work is well suited to automation precisely because it is rule-based, repetitive and requires no medical judgement.

Use cases from daily practice

Five use cases for GP and specialist practices

Every use case follows the same principle: the AI captures, sorts and prepares. Approving, signing and deciding remain with the practice.

AI phone assistant for routine calls

Starting point: at peak times many callers can't get through. What the system does: an AI phone assistant answers calls around the clock, identifies the request, asks for name, date of birth and callback number, and creates a structured task. If a caller mentions signs of an emergency, it clearly refers them to 112 or the out-of-hours medical service 116117. What changes: your team works through requests calmly instead of picking up the receiver every minute.

Preparing repeat prescriptions and referrals

Starting point: prescription and referral requests arrive by phone, email and handwritten note. What the system does: it collects requests across channels, matches medication and patient against existing data and bundles them for review. What changes: the doctor checks and signs the e-prescription as usual - just without the paperwork beforehand.

Sorting and assigning incoming mail

Starting point: results and letters land unsorted in several inboxes. What the system does: it reads incoming documents, recognises type, sender and patient, and suggests the assignment plus a short summary. Notable items are flagged, not assessed. What changes: less typing, fewer overlooked letters, a faster overview.

Practice knowledge on demand

Starting point: new colleagues keep asking the same questions. What the system does: an internal assistant built on a RAG system answers questions about your quality manual, hygiene plan, procedures and billing notes - citing your own documents. What changes: faster onboarding, and knowledge stays in the practice when someone leaves.

Patient information and reminders

Starting point: preparation notes for examinations, vaccination recalls and appointment reminders take time. What the system does: it turns your approved templates into clear information, in several languages if needed, and sends reminders through the channels you already use. What changes: fewer follow-up questions, fewer missed appointments.

As of October 2026

ePA, e-prescriptions and KIM: what the Telematics Infrastructure already requires

Over the past few years, Germany's Telematics Infrastructure (TI) has created facts that every AI solution must connect to rather than bypass:

  • E-prescription (E-Rezept): Since 1 January 2024, doctors have been required to issue prescription-only medicines covered by statutory health insurance as e-prescriptions. Patients redeem them with their health card, an app or a printout.
  • Electronic patient record (ePA): Since 1 October 2025, practices have been obliged to use the ePA, including uploading findings reports and doctors' letters. The medication list is generated automatically from e-prescription data, and after the card is inserted practices have access for 90 days by default.
  • KIM: The secure email service for German healthcare replaces fax and post for sending doctors' letters and results and can be integrated into the PVS.
  • DiGA: Digital health applications can be prescribed digitally since October 2025.

What does that mean for you? Structured data increasingly exists in digital form - but spread across PVS, KIM mailbox and ePA. An AI layer can help precisely at these hand-overs: pre-sorting documents, bundling tasks and summarising information for the team. The PVS remains the leading system. How deeply we integrate depends on the interfaces your PVS offers; we check this before every pilot.

No patient should be stuck on hold just because someone else needs a repeat prescription.

Safe on board

Data protection, confidentiality and the line to medicine

A medical practice processes almost nothing but health data. That's why we settle these points before writing the first line of code - documented, not just promised. Our full review framework is described under GDPR-ready AI.

Health data under Art. 9 GDPR

Medical confidentiality and data processing agreements

Hosting in Germany

No medical device, no diagnosis

How we work with your practice

No big bang, no PVS migration. We start small and build on what demonstrably works in daily practice.

  • 01

    1. Analysis: where does the time go?

    Together we look at which call reasons and document types dominate, which PVS and TI connector you use and which interfaces exist. The result is a prioritised list with effort and benefit per use case.

  • 02

    2. Pilot: one use case, real calls

    We start with a clearly scoped case, often the phone or prescription requests. Your team tests it in daily operation while we fine-tune dialogue, templates and hand-overs. Only when the pilot holds up do we move on.

  • 03

    3. Rollout and operation

    We extend to further cases, train the team and run the system in Germany. Updates, monitoring and adjustments to new TI requirements are part of operation, not a follow-up project.

From single practice to group practice

A single-handed practice often only needs a lean phone assistant with a task list. As soon as several doctors, locations or specialties come together, a shared foundation pays off: with Aipe, our GDPR-compliant AI operating system, phone, inbox and knowledge assistant run on one platform with role-based access rights. Reception sees only what reception needs, billing only what billing needs.

If you work in a larger network, also read our page on medical care centres (MVZ). How pharmacies handle the other side of prescription requests is covered on our pharmacy page.

Sailrs is an AI and software agency from Stuttgart, founded in 2019, with almost 150 completed projects. We work across Germany, and the initial call can happily take place by video between two consultations.

FAQ

Frequently asked questions about AI for medical practices

Can an AI answer patient calls in a German medical practice?

Yes, provided data protection and medical confidentiality are properly arranged. That includes a data processing agreement, a confidentiality undertaking by the provider, hosting in Germany and clearly informing callers that they are speaking with an AI assistant. The assistant records requests but makes no medical decisions.

Will the AI phone assistant replace my medical assistants?

No. It takes over recurring routine calls such as prescription requests, rescheduling or questions about results and files them in a structured way. Your assistants gain time for patients at the desk and for matters that need a human touch.

Does it work with my practice management system?

The PVS remains the leading system. Depending on the vendor, we connect via existing interfaces, import folders or task lists. Which option works for you is clarified in the analysis phase before a pilot starts.

Is AI for practice administration a medical device?

Not if it exclusively administers, communicates and provides knowledge. Software that makes diagnoses or recommends therapies can fall under the Medical Device Regulation. We deliberately do not build such functions.

What does AI for a medical practice cost?

Projects at Sailrs start at around 15,000 euros. A phone assistant with standard integration sits at the lower end; deep PVS integrations or several locations cost correspondingly more. We narrow down the scope in a free initial consultation.

How long does a pilot take?

That depends on the use case and the interfaces. We plan the pilot so your team can test the system in real practice operation before any decision on rollout.