AI Assistance in Recording Client and Patient Information

• Social and healthcare professionals spend several hours daily on documentation tasks. We aim to provide our professionals with more time to interact with clients.
• The purpose of the trial is to test whether AI solutions can save time spent on documentation and simultaneously improve the quality of records. Additionally, we want to understand if this can enhance the client experience and reduce staff workload. 

The AI application converts recorded speech into text and makes a draft based on this. This reduces the time professionals spend on documenting client and patient records. The conversation is recorded into short segments, which are automatically deleted after the draft is created. The recording is not stored as an audio or text files to which one could return. The recorded segments are immediately deleted once the AI has finished the draft.

• The recording is not stored as an audio or text files to which one could return. The recorded segments are immediately deleted once the AI has finished the draft.
• The professional documents the information into the client and patient information system, and you can access your documented healthcare information via OmaKanta, as before. Additionally, social services documentation will become available through OmaKanta by 2026 at the latest.

As this is a new way of producing documentation, we are proceeding cautiously. We have expanded the trial step by step, one service at a time.‌ The application is currently used by doctors, nurses, physiotherapists, occupational therapists, psychologists, and social welfare professionals.

The draft is visible only to the professional treating you.  

The information does not go directly into the client and patient information system. The professional is always responsible for their patient records.  Professionals will always review the draft carefully and make necessary corrections before storing it to the system.   

The application utilizes large language models, but the data flow is restricted to ensure that no personal data can be used for training language models. Anonymized corrections made by professionals to drafts may be stored in Gosta Labs Oy’s internal training database, but all identifying information is removed. This database contains only brief snippets of text, never full conversation transcripts. None of your information will leave Gosta Labs for external language model training. Personal data remains exclusively in the client and patient information system.  

Yes. At the start of the appointment, the professional will always ask for your consent to start the recording. If you refuse, the professional will document the information in the traditional manner. The professional is always responsible for the records entered into the client and patient information system. Participation in the trial or refusal to participate will not affect the service or treatment you receive.  

The Gosta application processes personal data when converting speech into a documentation draft but using the service does not store personal data on Gosta Lab Oy's servers or the professional's computer. The Gosta application does not maintain a client register. Client and patient data is maintained exclusively within the client and patient information systems used by the Western Uusimaa Wellbeing Services County.