Regtech solutions are increasingly using Artificial Intelligence (AI) such as machine learning and data analytics to bear on regulatory requirements and processes, making efforts to keep illicit finance out of the financial system more effective and efficient.
The Regtech opportunity
One of the big challenges facing banks today is extracting accurate insights from the huge volumes of data they process. Yet the answer isn’t to reduce the amount of data we process. Rather, Regtech enables banks to gain the necessary insights without necessarily increasing resources.
This means reduced likelihood of suspicious activity being overlooked, and a reduction in the number of ‘false positives’ flagged. Financial crime analysts can therefore focus their time on the higher-value work of investigating true matches and identifying useful financial intelligence for law enforcement.
In the medium term, Regtech is likely to offer augmented intelligence – assisting human decision-making, rather than replacing humans.
The impact so far
Regtech is already having an impact is on name screening – the process of checking whether individuals involved in a transaction are on lists of higher risk individuals, or subject to sanctions. It can also assist in transaction monitoring and investigations, though use in these areas is in the early stages.
Standard Chartered has been working with technology provider Silent Eight since 2018 to develop and implement such new technology. We aim to use machine learning and natural language processing to optimise name screening, which often generates large numbers of ‘false positives’ because of similar or identical names. As Martin Markiewicz, the CEO & Co Founder of Silent Eight says, the ambition is “to help revolutionise how screening will be done in the future.”
In a proof of concept trial – which covered name screening for adverse media, sanctions, and politically exposed persons – we removed 40 per cent of alerts the system currently generates as false positives. And analysts were able to process a further 30 per cent of the alerts generated more efficiently because the new technology had completed part of the investigation for them. This also meant that analysts made fewer mistakes: our estimate is that human error rates were reduced by 50 per cent.
Looking to the future
Successful development of Regtech solutions over the coming years will have three key characteristics:
Firstly, and most importantly, these solutions will need to be developed in partnership between the banks implementing them and the providers developing specific applications, engaging with regulators where appropriate. In our experience, taking a non-competitive approach to Regtech (by sharing our understanding and insights with other financial institutions) has been critical for the uptake of Regtech solutions. In July, the UK Financial Conduct Authority held a ‘TechSprint’ in London and Washington DC, bringing together financial institutions, technology providers, law firms, and others, to find better ways of increasing prevention and detection rates for financial crime. Standard Chartered worked with our partners to present two solutions: one to safely share data between banks to enhance due diligence, and another to more effectively share data using graph functionality.
Secondly, banks and regulators will need to work closely together, especially in instances where both existing methods and new emerging technologies in implementation phase are running in parallel. This is to be expected at first as industry and regulatory confidence grow. However, going forward, effectively addressing regulatory concerns early in the implementation stage will help avoid unnecessary ‘parallel runs’ and contribute to quicker and more efficient deployment of new industry solutions.
Thirdly, the industry recognises that technology is not ‘perfect’ or ‘zero-error’. Human analysts sometimes make mistakes, so an automated process with fewer errors – while not in itself error-free – would be an improvement. There is a need to agree within the industry on what margin of error would be acceptable, and for regulators to reach agreement on that too. Martin Markiewicz concludes “introducing AI into this space is not straightforward, with regulatory oversight and tolerance for errors extremely low.”
Looking to the future, it’s clear that Regtech solutions will play an ever-greater role in the fight against financial crime. However, this cannot work without collaboration: “We believe that financial institutions can both increase effectiveness and efficiency and that Regtech has a role to play,” says Markus Schulz, Standard Chartered’s Global Head of Financial Crimes Compliance Controls. But he cautions that, “the goal of achieving both greater effectiveness and efficiency cannot be outsourced to technology providers and there are no silver bullets though there are some quick wins.”
This article was also published in Bankable Insights Issue 11