November 28, 2018 / 1:30 PM / 12 days ago

HSBC set to launch cloud-based AML system next year, says senior official

LONDON/NEW YORK(Thomson Reuters Regulatory Intelligence) - HSBC hopes to launch a new cloud-based system for anti-money laundering (AML) risk management next year, a senior official said. The bank has been using data analytics in cooperation with Google Cloud as part of the development work, said Jennifer Calvery, global head of financial crime threat mitigation.

A sign at a HSBC bank branch office is seen in New York September 18, 2008.

The bank wants to use machine learning to analyse data from the activities of its more than 38 million customers to better help it spot suspected criminal activity, said Calvery, a former director at the U.S. Financial Crimes Enforcement Network. The new prototype system has shown positive early results.

The bank has engaged with regulators to explain its new system, Calvery told the Singapore Fintech Festival. Authorities in the Lion City had been notably supportive of having such conversations with the bank and exploring new avenues of financial crime risk management, she said.

Calvery outlined a system that could draw on big data to assign a financial crime score to each customer and single out those with a higher probability of committing crime.

“We are looking for criminal activity in a completely different way, which will require us to fundamentally change policies and processes,” she said.

“What we have seen when testing the prototype is that it is finding suspected criminal activity that had only been identified in the past by humans, not by our own systems.”

ENORMOUS PROCESSING POWER

The new system could analyse data from customer activity in the more than 60 jurisdictions in which the bank operated. The sheer processing power this required meant the only viable option was to seek a partnership with a cloud computing giant such as Google, Calvery said.

The system analysed data on customers’ behaviour, comparing it with the behaviour of other similar customers and flagging anomalies for review. The system also analyses customers’ transactions and their networks to establish a picture of who they are dealing with and whether there are any financial crime indicators in customers’ wider networks.

Earlier this year, HSBC said it would have around 100 petabytes of data on the Google Cloud platform by the end of 2018. Its two first projects on the cloud are focused on using machine learning for the AML prototype and for country-by-country liquidity reporting.

One of the hardest parts of managing the data coming out of the prototype system was what to do with the information, Calvery said.

“That’s the hardest part of building this thing,” she said. “We have to deliver that information to the right people at the bank — we have more than 230,000 employees — at the right time to take action. We have to have some level of oversight and governance to make sure that happens in the right way. That’s the piece that requires us to rethink our policies and procedures.”

The bank was in close communication with regulators to make them comfortable with the idea of moving to this system.

“I’m talking about taking what the financial industry does today, in terms of managing financial risk, and making it more efficient,” Calvery said.

“It’s a gamechanger because it’s something we can scale when you operate in more than 60 jurisdictions. It will enable us to have consistency in how we do AML.”

This article was produced by Thomson Reuters Regulatory Intelligence and initially posted on Nov. 16. Regulatory Intelligence provides a single source for regulatory news, analysis, rules and developments, with global coverage of more than 400 regulators and exchanges. Follow Regulatory Intelligence compliance news on Twitter: @thomsonreuters

0 : 0
  • narrow-browser-and-phone
  • medium-browser-and-portrait-tablet
  • landscape-tablet
  • medium-wide-browser
  • wide-browser-and-larger
  • medium-browser-and-landscape-tablet
  • medium-wide-browser-and-larger
  • above-phone
  • portrait-tablet-and-above
  • above-portrait-tablet
  • landscape-tablet-and-above
  • landscape-tablet-and-medium-wide-browser
  • portrait-tablet-and-below
  • landscape-tablet-and-below