Doctors spend 35% of their time on administrative tasks which can be better spent on actually treating patients. Additionally, most doctors are currently not filling in the structured fields on the Electronic Health Record (EHR), resulting in a huge loss of data. An incomplete patient record can cause problems when a patient is referred to a specialist doctor and could even lead to incorrect diagnosis.

Autoscriber offers a software solution that aims to save the physician 30% of their time they spend on administration while more accurately extracting all the important information. This is done by listening in on the doctor-patient consultation and transcribing this using Speech-to-Text models. From this transcript, all the important information mentioned in the consultation like symptoms, medication and procedures is extracted using trained Artificial Intelligence models. This structured data can be automatically populated into the EHR and used for diagnosis support.

At present we are collaborating with three hospitals in the Netherlands to validate and improve our software. In order to develop this further we require EUR 2.4 million which will finance software development costs, onboarding of additional hospitals, annotation of conversations to improve model performance, and compliance certification (ISO and NEN). Our core team comprises award winning researchers in Natural Language Processing (NLP), an experienced co-founder of successful start-ups, AI professionals, and annotators with medical training to ensure quality data labelling to train our AI models. We are supported by a strong network of AI and medical professionals as members of the Thematic Technology Transfer - Artificial Intelligence network.

Jorijn Enterman
Liza King
Koen Bonenkamp
Jacqueline Kazmaier
Startup activities

Venture Challenge Spring 2022

Jorijn Enterman