A recent NLP innovation may hold the key to better financial crime fighting. By creating complex, multi-dimensional relationships among words, text embeddings technology transforms the semantic relationships within languages into mathematical expressions. While this innovation has revolutionized many familiar Natural Language Processing (NLP) applications like translation, its potential extends far beyond the familiar. The same mathematical approach used to identify semantic similarity across languages could also be used to spot fraudulent bank activity, giving financial institutions a new and powerful approach to fighting crime.