Chatbots accomplishing human-like expertise in the domain of legal insurance

Abstract

Chatbots and digitisation were a frequent topic thorough the year 2017. Different branches and business sectors get captured more and more by cost pressure and need for innovation. So, does the insurance industry. Within this paper, the limited nature of current chatbots is examined. To effectively understand state of the art models, different techniques in machine learning and case based reasoning were analysed. The aim was to examine whether it is possible or not, to design a chatbot, that accomplishes human like behaviour and expertise in the sub domain of legal insurance. Using the Design Science Research in Information Systems (DSRIS) approach, a chat-bot was developed and tested over several iterations. By evaluating use-case complexity aiming to create significant results, the process of a “shallow audit on coverage” within the legal insurances branch was targeted. Findings derived from literature were applied to create a prototype, which was utilised to experiment and process observations. The model, as final artefact resulting from software development, is illustrated below....