Machine learning technologies are only as good as the amount of data they can ingest for the learning phases. One of the characteristics of threats like spear phishing emails is that they are very low volume. There is therefore not enough data to train a machine learning engine with great accuracy. Vade has developed and patented a technology for augmenting textual data to generate copies from originals to accurately train threat detection models using machine learning. This technique is the basis of the anti-spear phishing engine currently used within the Vade for M365 product.