Compliance is a function of growing importance within banks and financial firms, providing a (partial) guarantee that regulations are respected, and that institutions retain their licences to operate.
Following a rise in prominence and a period of structuring, compliance is on the verge of a significant shift. The function is about to become a lot more efficient and effective, thanks firstly to new technologies, but also – and above all – humans.
While technologies such as robotic process automation (RPA) and natural language processing (NLP) are now common in banks’ compliance processes, the use of artificial intelligence (AI) has not yet reached that stage. Indeed, in the context of market abuse prevention and the fight against money laundering, the trade surveillance mechanisms currently in use are based on an approach that determines scenarios using static rules.
Beyond banking, AI plays an increasingly important role in our lives, simplifying many things we do. Examples include cars that alert the driver when a lane boundary is crossed and rectify the situation by moving back into the correct lane. Another example is an email inbox that suggests increasingly relevant automatic responses to messages.
Imagine this applied to banking compliance, and more precisely to anti-money laundering and the prevention of market abuse. As an example, systems based on neuron networks could slash the rate of false positives (which are ridiculous at present), and perform an autonomous analysis of valid cases.
Big data could enable AI to go even further. The ability to process ever larger batches of data as well as retrieve the desired information and reproduce it in the most relevant and legible manner, will signal the arrival of a new era. This era is not far away, as evidenced by many factors, including the transition from data warehouses to data lakes, the intensifying debate surrounding in-house data storage versus cloud storage, the enthusiasm of fintechs and IT managers, and the recruitment of growing numbers of data scientists by financial institutions.
In addition, Google and social networks are becoming an increasingly important tool compliance offers use for client research. Should we deduce from this that traditional tools have had their day? If the answer is ‘yes’, the transition to an entirely new model will need to be managed consistently and with appropriate support.
However, financial institutions are reluctant to expand their use of big data systems due to IT security concerns. Any new infrastructure or solutions will need to offer at least the same level of security as those currently in place, and this is not yet taken for granted.
THE HUMAN ELEMENT
This is the most complex part of the compliance equation, given the potential implications for humans. After all, we’re already witnessing robotization affecting jobs in finance.
Today, verification systems that use new technologies leave the final step to humans. If this changes in the future, what will the regulator’s position be and who will audit the algorithms put in place? Will it be humans, robots directed by humans, or autonomous robots?
To progress, firms and regulators must jointly define a clear framework for future compliance, where man and machine co-exist and enhance each other’s roles. There’s still much work to be done to make that happen.