Munich Fire Department Implements AI Voicebot for Non-Critical Patient Transfers
Munich, February 27, 2026 – The Munich Fire Department has announced the implementation of an AI voicebot to manage planned, non-critical patient transfers. This new system aims to alleviate the burden on the integrated control center, ensuring that emergency calls to 112 continue to be handled by human operators without delay. The initiative was presented at the Microsoft AI Tour.
Addressing Capacity Challenges in the Control Center
According to the Munich Fire Department, its integrated control center receives approximately 2,500 calls daily, with around 400 of these being requests for patient transfers. Emergency calls always take precedence, which can lead to waiting times of several minutes for coordination between the control center and hospitals regarding planned transfers.
This is where the new project, developed in collaboration with Microsoft Germany, comes into play. The AI voicebot is specifically designed to handle planned, non-critical patient transfers, such as transfers between hospitals. The pilot phase demonstrated that the bot efficiently captures all relevant information during the initial intake process. This automation is expected to reduce the average processing time per call by 3.5 minutes, previously handled by human staff.
Clear Division of Tasks and Robust Safeguards
A critical aspect of the new system is the clear division of responsibilities: emergency calls via 112 will continue to be handled by trained personnel in the integrated control center. The voicebot operates with a multi-stage security system: it requests mandatory information, then verifies the details and location against databases. Finally, trained dispatchers check the plausibility of the request and allocate appropriate rescue vehicles, ensuring that critical decisions remain with human experts. If the AI detects any uncertainties or a potential emergency, the call is immediately transferred to a human call taker. The same applies to any discrepancies in the patient transfer request.
The bot’s capabilities are clearly defined:
- It only handles planned, non-critical patient transfers.
- It does not influence vehicle allocation or provide waiting times.
- It cannot cancel or modify transfers.
- A fallback to a human dispatcher is always possible if needed.
Technological Foundation and Data Security
The solution is based on Microsoft Azure technology, incorporating Azure AI Search for location verification and HD Voice technology. Data protection and IT security were paramount from the outset. Call acceptance and data storage occur within the Munich Fire Department’s environment, ensuring compliance with relevant regulatory requirements, including the European AI Act and BSI guidelines.
Successful Pilot Phase and Future Rollout
The voicebot was tested at the LMU Klinikum Großhadern, whose emergency department treats up to 150 patients daily, with 40 to 50 requiring inpatient admission. Professor Dr. Matthias Klein noted that delayed transport registrations could lead to patients being treated but unable to leave the emergency department. During the test phase, Klein observed that transfers could be registered immediately, missing information was specifically requested, and there were no waiting times. The AI voicebot is expected to significantly relieve the integrated control center of the Munich Fire Department regarding patient transfers.
Following a test operation in February, a multi-stage rollout is planned for the LMU Klinikum Großhadern in spring. The project aims to gradually expand the solution to other hospitals in Munich and subsequently across Bavaria. Initially, the voicebot will be available to hospitals and doctor’s offices via a special number; the existing transport number 19222 remains unchanged.
Fire Director Wolfgang Schäuble justified the move by citing increasing call volumes and limitations in personnel and space. “Automation was a necessity for us. One could say: AI helps – humans save.”
Broader Context: AI in Emergency Services
The topic of “AI in control centers” has previously been covered in Feuerwehr-Magazin, focusing on live translation for language barriers in the Bayreuth/Kulmbach integrated control center. The Munich approach, however, focuses on a different area: the structured acceptance of planned patient transfers to relieve the control center, rather than translation in emergency calls.
