AI and cloud combining to fight fraud - Temenos

Artificial intelligence (AI) technologies are being directed to detect financial crime, according to banking software firm Temenos.

AI techniques are teaming up with traditional analysis to work on ever greater volumes of data, which are now being stored on cloud services.

“In financial crime mitigation we’re focusing on AI methods mainly to detect and prevent fraud. In the space of anti-money laundering and sanctions compliance regulators are rather reluctant to the use of AI and still prefer rule-based solutions,” said Marlene Meli, head of compliance practice at Temenos.

“Rule-based fraud detection is being combined with AI, following the money flow but also looking at other signs that can be a hint that something strange is going on, for example through online monitoring who is accessing what, at what time, and by which channels and devices,” she said.

“We like to have a combination of rules-based detection and AI. The AI is being used to detect hidden patterns and to derive additional information, as a sandbox in which to test new techniques,” said Meli.

The AI is good at going through the mass of historical data, increasingly kept on cloud services.

The firm is working with a mix of large financial institutions to smaller regional banks, noted Adam Gable, product director for financial crime mitigation and treasury at Temenos.

“We’ve been investing around AI and robotic process automation (RPA) to speed up the processing of alerts,” said Gable.

The firm is using its own proprietary AI capability, he noted, developed by its data and analytics division.

“AI has broader capabilities, which we are capitalising on” said Gable.

Meli suggested more firms are taking a collaborative approach to tackling fraud.

“For collaboration and data sharing, where data protection laws permit, the cloud is a trend,” she said.

Gable added: “Cloud is one of the pillars of Temenos’ strategy to meet the needs of today’s digital banks and financial services landscape. This is driving investment in software-as-a-service (SaaS) and platform-as-a-service (PaaS) capabilities as part of our Cloud offering. We have seen a 400% growth in cloud deployments between 2015 and the end of 2017.”

Anonymising data is still crucial for data sharing, she stressed, relevant for firms looking to use the cloud collaboratively to maximise the data from which AI can find fraudulent patterns or anomalies.

“Regulatory barriers are still one of biggest barriers they’re facing,” she said.

The cultural aspects of internal fraud are still vital, she suggested, highlighting the increasing number of internal fraud cases, including Volkswagen’s unethical behaviour in manipulating emissions tests, and the US retail banking arm of Wells Fargo, which created millions of fake accounts to meet sales targets.

“A get-tough-approach to boost performance creates a climate of fear which leads to a blame culture and unethical behaviour. . Senior management accountability, also from a culture and ethics perspective, remains important,” said Meli.

“Culture should cascade down, or in the opposite case you can have a top-down cultural failing,” she added.