A significant, and likely dominant, proportion of fraud is now conducted online. Police struggle to integrate this emerging reality into their processes, while expectations of this institution are high. These types of cybercrimes alter the volume and complexity of problems compared to how they previously manifested and require profound transformations of crime analysis methods to address them proactively. These developments face many difficulties, such as the quality of accessible data, the lack of existing analytical models and the need to increase police knowledge of fraud mechanisms within every level of organisations.We suggest methods to overcome these obstacles, which consist of implementing an approach integrating theories from various fields in criminology and forensic intelligence to examine the digital transformations of certain criminal processes. We take, as an example, how a generic script expressing the anatomy of an existing type of fraud can be used to interpret their new digital forms. This modelling activity both provides new insight into specific frauds and highlights relevant dimensions that are useful in orienting the development of crime analysis systems.