Our team worked hand in hand with the client virtually to provide a solution that significantly improved the client’s existing imaging and verification software. With our ML model, we could flag potential frauds among the millions of checks being processed every month. Moreover, adopting a neural network to parse a historical database of previously scanned checks, including fraudulent ones, allowed us to analyze scanned images of handwritten checks including variable elements such as payee, check number, account and routing numbers, signatures, etc. In addition to that, our AI experts taught the network to identify what was normative for good checks and what checks are anomalous.