HSBC has launched a global analytics platform that identifies potential financial crime by contextually analysing customer, transactional, and publicly available data in order to understand a customer’s global network. The platform, developed with AI solution provider Quantexa, illustrates how firms are deploying RegTech solutions to address the ongoing regulatory focus on financial crime compliance, with the size of financial penalties and the volume of reprimands reaching new highs.
HSBC’s Global Social Network Analytics (GSNA) Platform aims to strengthen the bank’s ability to use data more effectively to detect financial crime such as money laundering, terrorist financing, human trafficking, and bribery and corruption.
Struggling to combat financial crime effectively
HSBC operates across developed and emerging markets, including Asia-Pacific and Latin America. These are the kind of markets that can be the most challenging to monitor from a financial crime perspective. The bank’s ambition is to be industry-leading in its ability to combat financial crime, and is investing in its surveillance systems accordingly.
According to Richard Stocks, US head of financial crime at Quantexa, siloed data and transactional monitoring systems that cannot work with big data hamper the efforts of firms to detect criminal activity. “Traditional methods allow for too much risk,” he says. “They miss too many things, such as false negatives. An industry that talks about false positives often forgets that the missed risk is where the burden of financial fines hits.”
However, Stocks says false positives are a significant issue as well. “The sheer volume of false positives creates an enormous burden on the investigative units and the analysts downstream,” he adds. This volume of false positives can lead to ‘investigative fatigue’, he says, adding to the problem of false negatives as true financial crime slips through unrecognised.
To date, financial crime units at banks seeking to achieve the levels of compliance required of them by regulators have often had to add more analysts and investigators to sift through positive transactions, identify the ones that are more likely to be criminal, and submit suspicious activity reports.
Taking a new approach
This has meant that firms have struggled to effectively scale their financial crime activity – working within current approaches has just made it too challenging to accurately identify risk. Firms have been looking to AI, machine learning and other technologies, hoping that recent developments may help solve these financial crime detection challenges.
Quantexa offers an AI-based solution that takes a different approach to the way the data involved in the process is compiled and analysed. “The difference is when we look at potentially every other data quality matching tool vendor in the marketplace, they typically follow an order of operations data quality process for entity resolution in batch,” says Stocks. “The distinction is that Quantexa does all of that dynamically. You’re not forced to persist the data. It’s all done in memory and on the fly, so that any new pieces of information and insight might in fact change the outcome. It’s a multi-step process. And it’s always point-in-time.”
Quantexa simultaneously builds a profile and behaviour, including activities, transactions and attitudes in order to get to a mathematical single view. This can be determined based on weights to be in line with an organisation’s appetite for risk. Stocks says: “When you take this dynamic entity resolution approach, with multiple data points and a corroboration and conflict approach, the data quality problem goes away.”
Quantexa’s AI solution can be used to assess client behaviours drawing on data from multiple sources, including traditional documents, browsing behaviour, shopping behaviour, ATM usage, bank transfers and others.
Overall, this approach, Stocks says, can reduce financial crime risk and compliance risk, and creates the possibility to seize new opportunities. “If I can manage risk at a financial institution, it’s going to enable me to potentially enter riskier markets with riskier products for potentially riskier clients,” he says.