The AppZen Platform combines three critical capabilities: computer vision, semantic analysis, and deep learning. Our computer vision extracts text and images from unstructured documents associated with spend transactions such as receipts, invoices, and contracts. Our semantic analysis combines this extracted information with intelligence gleaned from business systems and external sources and develops a rich, contextual understanding of each transaction in our purpose-built data model that has been optimized for the finance domain. Finally, our platform’s deep learning continually refines its understanding based on billions of spend transactions consisting of trillions of data points, coupled with user feedback from auditors and reviewers interacting with the platform.