By PAUL O’DONOGUE, Senior Correspondent
PARTNERING with Google on AI helped HSBC revolutionise its AML system, according to a senior figure at the UK-based lender.
Nish Ranatunga, the head of financial crime risk models at HSBC, said its new AI anti-financial crime model has been a ‘game changer’.
“We’ve found more than twice the number of SARs (suspicious activity reports) for financial crime that we were finding in our [old] rules-based system. We’re doing it four times faster,” he said during a panel event at the ‘International Anti-Financial Crime Summit’ (IAFCS) 2025.
“That’s not just because of the machine learning capability. For us, it was really the decision to invest in the technology. And also open the doors to create optimizations.”
Google said in 2023 that it was working with HSBC on an AI-enhanced AML system. However, Ranatunga said the bank has been examining the technology for years before that.
“I joined HSBC about eight years ago. When I joined, I would say we were in the first phase of our AI journey,” he said.
He said the first phase was based on a rules-based system used by banks. This is where they were encouraged by regulators to closely follow official guidelines. However, a ‘rules-based approach’ has been criticised by many analysts for not properly prioritising financial crime risks. Bodies such as the FATF instead emphasize a ‘risk-based approach’, focusing on key threats.
HSBC and Google training AI model
Ranatunga said this was also an issue when training the bank’s AI model.
“The first phase was – taking the rules-based system, and creating some level of intelligence to score outputs,” he said. “To determine whether or not we want to send that to a human to investigate those results.
“A lot of us are probably in that area, in banking, where you have this very mature, long-running and inefficient rules-based system. And you create some kind of mechanics on top of it.”
Ranatunga said Jennifer Calvery, HSBC’s global head of financial crime threat mitigation, pushed for a more efficient approach.
“So to be able to move from a rules-based system that has some kind of adjudicator on top of it, to taking the technology that’s in that adjudicator. And making that the core detection platform – [that’s] where we are today.”
He said this was when the firm started working with Google.
“We partnered with Google. We helped them design and build an application to detect financial crime, specifically money laundering and [financial crime] risks,” he said.
“We use that as our core detection platform today across all of our transactional banking. That was a journey of about five years, really, from its inception.”
Ranatunga said it was important to give researchers time and space to develop the project.
He said HSBC and Google created a “space for innovation within our financial crime program”.
“Allowing them [researchers] to fail, and fail immediately. Because it took a lot of iterations for us to be able to build that platform effectively and responsibly.
“And we’ve now been using it for a number of years within many of our IT markets. [The system has] been mature for three or four years. We’ve reduced false positives, and we’ve seen a really, really big gain
We hope you enjoyed reading this article
If you would like unlimited access to AML Intelligence premium articles, newsletter delivered twice a week, access to our Global Bank Fines and Penalties database, free access to Boardroom Series events and much more, select one of our subscription options and become a subscriber!
