Published : 27/02/2026 - 3 minutes read
Customer Success Stories
UZ Gent Pioneers AI in Medical Administration
The Data Science Institute (DSI) at UZ Gent has developed an AI application that generates customized summaries of medical discharge letters tailored to the specific needs of healthcare providers.
Inetum supported the project by building a Microsoft Azure cloud environment that aligns perfectly with performance, security, and governance requirements.
As one of Flanders' largest hospitals, UZ Gent manages 7,000 employees and over 3,000 daily consultations. The DSI oversees high-quality medical data from electronic patient records to facilitate innovative research and policy-making.
Summarizing Complex Discharge Letters
Through the FRAIT project (Framework and implementation of AI tools), UZ Gent collaborated with several regional hospitals and GP associations to explore the potential of Large Language Models (LLMs). The goal is to summarize discharge letters based on professional preferences—for instance, a gynecologist requires different details than a general practitioner.
Cloud Governance and Personalized Prompts
While UZ Gent traditionally relies on on-premise infrastructure, Inetum designed a dedicated Azure cloud environment for this pilot, ensuring data remained within the EU. The research utilized GPT-4 and Databricks to develop personalized prompts. To ensure privacy, the study was conducted using realistic, anonymized data.
Exceeding Expectations
The evaluation showed that 70% of participating healthcare providers rated the summaries as 'good' to 'very good'. Formal accuracy reached 88%, while content accuracy stood at 78%. "The test demonstrated how we can refine prompts for even better results" says data scientist Mieke Deschepper. The long-term goal is to integrate this AI directly into the software used by physicians, allowing them to request summaries within the electronic patient file.