AI in Patient Care: Automatic Documentation with Noa Notes

Germany’s healthcare system is undergoing a digital shift - not just since the launch of the electronic patient record. Alongside growing IT standardization, AI now offers a way to automate time-consuming admin tasks. One major hurdle to patient-centered care remains: medical documentation.

Driven by legal and regulatory requirements, documentation has steadily increased. In hospitals, it takes doctors and staff up to 2.9 hours a day*; in private practices, around 1 hour daily or 7.5 hours a week.**

Inspired by ChatGPT’s breakthrough in late 2022, jameda explored how to use AI to reduce the documentation burden in healthcare. The result: Noa Notes, developed from an internal hackathon and launched in September 2024.

The assistant records medical facts during consultations and structures them individually. For reliable data transfer, jameda integrated buffering technologies from its video consultation tool. Whisper, an open-source speech recognition system, was additionally trained on 50,000+ hours of medical terminology and pharmaceuticals to ensure a accurate transcription and filtering of irrelevant information.

Noa Notes quickly proved its impact: In tens of thousands of sessions, average documentation time was just 4.5 minutes - providing the documentation directly after the consultation and saving 50–75% of the time depending on specialty.