With an increased focus from regulators and financial institutions on compliance, artificial intelligence (AI)-based solutions are being increasingly examined for their potential. One of the more interesting solutions is the application of natural language processing, which is capable of combining the results of surveillance and conversation monitoring to produce suspicious activity reports. In fact, natural language processing solutions have been tested successfully at a number of financial institutions. Despite this initial promise, there are still a number of issues to overcome if natural language processing solutions are to become viable options on a large scale, including the limited AI knowledge of financial institutions’ IT staff and challenges with integrating such solutions with existing legacy systems and cultures. This leads to uncertainty as to whether such investments will provide the return on investment necessary for many financial institutions to undertake them, especially more risk-averse financial institutions. That being said, the only way for such solutions to really develop within the financial services industry is for financial institutions to gain the capacities necessary to understand them and develop them. Only in this way will the real value of such solutions become a known and measurable quantity.