Being reachable by phone, coordinating appointments, transmitting lab results – for many medical practices, this is a daily state of emergency. This is precisely where [the company] comes in. Sinalis The Munich-based early-stage startup is developing an AI that automates patient communication, primarily over the phone. (See the videocast.) Pitch & People explains co-founder Paul Sachse, how a hackathon idea became a healthtech startup that aims to save time for medical practices.
"We handle communication with patients over the phone. AI accompanies patients throughout their entire journey: from initial contact and appointment preparation to follow-up care. Calls are filtered, relevant information is gathered in advance, and routine calls, such as those for transmitting lab results, are automated. The goal is to significantly relieve the burden on medical staff and better prepare doctors for their short appointments."
Originating from a hackathon – without a healthcare background
Sinalis was founded after a hackathon organized by TUM.ai, the student AI association at the Technical University of Munich. The team members initially met briefly via Discord, only meeting in person on the day of the hackathon, and after intensive discussions decided to focus on the healthcare sector. None of the founders originally came from the healthcare field. The decisive factor was the desire to develop a product with real impact. Initial user interviews with doctors followed that same evening.
Multilingual, natural, (almost) human
A key unique selling point of the Sinalis AI is its linguistic diversity. It communicates in over 70 languages, including Turkish, Arabic, Ukrainian, and Russian – an important feature for medical practices that regularly encounter language barriers. According to interviews with doctors, this affects around ten percent of their patients. Feedback from initial tests has been surprisingly positive. Some patients didn't even realize they were speaking to an AI, while others curiously inquired about the origin of the voice.
Technically, the system is based on a combination of different speech-to-text and text-to-speech models. Its real strength, however, lies in the fine-tuning: tone of voice, response length, and empathy can all be adjusted. While some patients prefer concise communication, others desire more empathy. This is precisely where the team is currently conducting intensive experimentation.
Between data protection and everyday practice
Sinalis is still in the pilot phase. Alongside product development, the team is working on data protection and regulatory issues – a "jungle," as Sachse describes it. All services used must run on EU servers, and it's also crucial to clearly define when a product qualifies as a medical device. For the time being, Sinalis is therefore deliberately focusing on organizational processes and refraining from making medical recommendations.
One clear goal: more time for patients
In the long term, Sinalis aims to do far more than just automate phone calls. The vision: a central communication interface for doctors' offices and, later, hospitals. Applications in telemedicine or the integration of external devices are conceivable. In the short term, however, the focus is on a very concrete problem: lack of time. Medical staff have, on average, only a few minutes per patient – insufficient for comprehensive treatment. Improved preparation and follow-up care are intended to change this.
So far, the startup has financed itself with the €10,000 prize money from the TUM-ai Hackathon. Sachse and his colleagues are currently applying for funding programs and preparing for investor pitches. But regardless of the financing, for Saxon One thing is clear: personal motivation is impact.
"When I think of my grandmother, I realize how little care she receives in the medical system."
This is exactly where Sinalis wants to start – with technology that works in the background so that people in the foreground have more time for each other.
