Insurance fraud is a growing problem. This article describes a project that aimed to specify and develop a suite of computer-based tools to support the early detection and subsequent investigation of potentially fraudulent claims. System requirements were informed by ethnographic studies focusing on: (1) understanding current fraud detection practices and (2) characterising fraud investigation expertise. Tools were designed that sift claims for potential problem cases and assist in the processes of investigation and detection of new fraud types by providing anomaly capture, argumentation and visualisation environments. The resulting tools capitalise upon expertise and embody processes that can subsequently be used by inexperienced claims handlers to detect and deal with fraudulent claims.