Why Hasn’t The Innovation Train Arrived to Fincrime Compliance?
Challenger banks are using innovative technology to change the game with customers. From the use of open banking, which is dramatically improving customer engagement and creating frictionless onboarding, to the creation of smart products and tools to make banking more personalized. Established banks are jumping on the innovation train for customers and following suit. But are back-office solutions making the same leap forward on innovation? It is widely reported that financial crime risk management continues be more and more costly and challenging. It is time to stop the inefficient and ineffective ways financial institutions carry out anti-money laundering (AML) and financial crime reviews.
Rules-based approaches hold sway, staffing has been increased to meet regulatory requirements and reported costs are exponential. The challenge to take a different approach is exacerbated by organizations only having access to their own data on entities and people. Data sharing is limited, which also limits the ability to prevent financial crime from occurring across organizations. For several reasons, a bad actor that is known at one bank cannot be seen by another bank. Even the largest banks struggle with insufficient data to carry out effective Customer Due Diligence (CDD). Without adequate broad customer insights, there can’t be an adequate risk assessment of that customer. Challenger banks are particularly exposed to this environment as they rush to build their customer bases.
In traditional banks, the approach seems to be to add more to existing processes to solve the problems. Where is the innovation that is so cherished to attract and retain customers but is a second thought in protecting society? It is not as though the current operational approach is under-invested. What solutions should we seek?
A recent review found that some challenger banks in the UK failed to sufficiently check their customers’ identities, even though they are using tech for verification purposes, such as matching passports to live photos. However, identity manipulation is relatively easy. In my opinion, the challenge then becomes improving how the bank assesses financial crime risk. Transaction monitoring and CDD processes are expensive and inadequate. Rather than suggest challenger banks use older, more expensive, less efficient solutions for CDD, we should consider more effective and innovative anti-money laundering (AML) solutions.
Data sharing could be the answer, but this is fraught with privacy, security, legal and technology risks. How would it ever be possible to share enormous amounts of data considering these challenges? Wouldn’t it be amazing if the industry could find a way to share insights around customers’ data without violating privacy laws so that the financial infrastructure could take a shared responsibility to prevent money laundering and financial crime? It’s time for a different, more effective way to manage customer risk.
Federated machine learning models could be the answer, allowing new banks to benefit from the experience and knowledge of older and more established peers. Sharing these models that learn and evolve as they are used at different organizations is one answer. It is innovative and compelling in that no Personally Identifiable Information (PII) moves or is shared. The lack of data at a new bank wouldn’t mean it is at a disadvantage in fighting financial crime. It would be able to rely on shared insights and behaviors surrounding secure customer data sets to meet and exceed regulatory needs despite its newness. It doesn’t only benefit the challenger banks. It also benefits the established players as they get to see entire local and even global market behavior.
Federated machine learning is new. It solves existing problems and brings a new dimension to efficiency and effectiveness for all financial organizations.
