- Research Project
Virtual research assistant Artificial intelligence for lawyers
A virtual research assistant based on artificial intelligence (AI) and focused on medical insurance law allows lawyers to find relevant information up to ten times faster and more precisely than ever.
- Lead school School of Engineering and Computer Science
- Institute Institute for Data Applications and Security IDAS
- Duration (planned) 01.11.2019 - 30.04.2022
- Project management Erik Graf
- Head of project Erik Graf
- Partner Innosuisse
- Keywords Insur-Tech, Legal-Tech, Machine Learning, Innosuisse, Start-Up
Insurance companies and law firms handle numerous complex cases that often involve extensive research work. Processing and analysing these cases manually is time-intensive, and working through complex case dossiers that are sometimes thousands of pages long creates a heavy workload for the parties involved.
In conjunction with the InsurTech startup legal-i, researchers at BFH have developed a virtual research assistant that is based on artificial intelligence (AI) and focuses on medical insurance law. Adaptive machine learning approaches were used to create an intelligent system that is able to analyse complex legal and medical issues. Its development stems from a close collaboration between BFH researchers and the Bern startup. The efficiency of the solution in productive applications will be optimised in a series of collaborations with Swiss insurance companies and legal firms, using legal-i’s pilot projects as a basis.
Specialists in medical insurance law can use this AI to find relevant information up to ten times faster and more precisely than ever. The AI-based software is able to make handwritten and scanned documents searchable, as well as digitising and sorting them. This allows specialists to focus on tasks that they perform better than machines (e.g. analysis, argumentation, strategy development and decision-making). The system’s capabilities will be gradually expanded over the course of the project to enable it to recognise case-specific facts in a targeted fashion and prepare them for further processing.