Technology is an essential weapon for financial services firms in the battle against anti-money laundering (AML). The Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation, the Financial Crimes Enforcement Network (FinCEN), the National Credit Union Administration, and the Office of the Comptroller of the Currency (OCC) all recognize this and issued a joint statement earlier this month encouraging financial institutions to try new and innovative ways of combating AML. Three types of technology – advanced analytics, software robots, and artificial intelligence (AI) – could help make it easier to detect and prevent money laundering, as well as comply with existing regulations and follow the guidance of this joint statement.
Advanced analytics are the most frequently used of the three technologies today. They can take the form of more sophisticated dashboards, reports, alerts, and scorecards. Examples include:
- AML model validation and tuning – For understanding and presenting the results of model validation of BSA/AML and OFAC systems as well as other models.
- Data profiling – To analyze and help firms understand the data from core systems by collecting relevant statistics such as frequency or distribution analysis. Results from such analysis will help firms optimize and tune their detection models.
- Data quality analysis – To analyze and track the quality of the data that is fed into AML systems. This can also be used to assess the impact of poor data quality in the models.
- ATL/BTL testing – To help perform above the line (ATL) and below the line (BTL) testing for BSA/AML, OFAC, and customer risk models to evaluate detection coverage and model performance at the firm.
- Threshold analysis – The use of what-if analyses in models to fine tune and enhance effectiveness and efficiency.
- Capacity planning – To understand how much time is spent on case investigations and perform staff planning based on case and alert volumes, productivity, and other factors.
- Metrics and Reporting – To obtain real-time metrics such as efficiencies in case management, changes in risk and for standard reporting.
These types of approaches can improve the quality of the information that executives use when making decisions and speed up the decision-making process. Advanced analytics can also provide a more unified view of the AML program across the whole organization.
Called robotic process automation (RPA), software robots are capable of performing a range of tasks on existing applications. To be successful, this technology needs to be applied to repeatable, rule-based or logic-driven tasks. While the industry does talk about so-called “cognitive RPA” and “intelligent RPA”, RPA applications currently used for AML program purposes are relatively basic. Use cases currently found in firms include:
- Repetitive processes within model validation and tuning – Back-testing, ATL/BTL testing, and sampling can be automated using RPA.
- Data reconciliation – Information from one system can be compared with data from another system automatically via RPAs.
- Repetitive tasks within AML case investigations – RPAs are great for high-volume, low complexity tasks.
- Customer onboarding/KYC – Software robots can be combined with optical scanning and recognition technology to automate paper document processing and loading those documents into digital customer files.
- Gathering case research information online – Robots can be set to search the web for individual names, company names, negative news, and other terms to provide to the investigator when the case is opened.
Deployments of RPAs can lower both costs and errors within processes. They are usually relatively inexpensive to create leading to a rapid return on investment. Lastly, the activities of an RPA are auditable, which is particularly valuable within an AML compliance setting.
AI has the potential to make a tremendous difference within the AML space by introducing in the ability to identify complex vulnerabilities, threats and patterns that current processes cannot detect. It can also minimize the volume of false positives that current systems produce. Large financial institutions and Regtech firms are currently engaged in pilot AI programs that will eventually augment or replace existing systems. However, there has been limited progress in the adoption of AI technology.
Forms of AI that may be familiar include machine learning, pattern recognition, natural language processing, neural networks, and chatbots. The four types of AI are:
- Purely reactive – This is the most basic form of AI. It simply reacts to what it encounters. However, it can learn over time to improve outcomes. AI is also capable of working with both structured and unstructured data.
- Limited memory – In this form of AI, the computer considers past information and adds it to pre-programmed representations of the world. In AML, limited memory AI could be used in case research disposition – that is, making decisions on cases. It could also be used in KYC due diligence to develop contextual awareness on documents by going through a number of documents and applying what it has learned in the past to aid humans in understanding materials more quickly.
- Theory of the mind – This is when the AI technology has the ability to understand human thoughts and emotions and can interact with people on that basis. This technology is theoretical, however, it could be a very powerful tool for AML purposes when it is available. For example, it could predict suspicious behaviour with a higher degree of accuracy than humans are currently able to.
- Awareness about itself – This level of AI, extending theory of the mind, can understand its own feelings as well as and the thoughts and feelings of those people it interacts with. Again, this technology is not yet available, but if it were, it could perhaps replace humans and even prevent suspicious activity from happening in the first place.
Deployment of even the first two basic levels of AI could result in cost and error reduction. AI can also be deployed over a large scale to streamline workflow. Its flexibility, and ability to make decisions on the fly, could potentially give it a more proactive role in AML detection and prevention.
In conclusion, in the fight against money laundering, a range of new technologies are already available to firms while the future looks bright for the ability of evolving technologies to help firms detect and combat AML