AI Is Only as Good as the Datasets They Access, Says Deutsche Bank

As the world becomes more interconnected, cross-border payments have never been more important in the banking and finance industry. 

However, the current system can be slow and cumbersome, making it difficult for banks and their customers to process global cross-border payments efficiently. 

That’s where the value of ISO 20022, a new international standard for exchanging electronic messages which all global banking and financial service actors will have to adopt by 2025, could help fill in the gaps, said Ole Matthiessen, global head of cash management at Deutsche Bank. 

In a recent interview, Matthiessen highlighted some of the benefits financial institutions (FIs) and their corporate clients stand to gain from the new messaging system, including access to a more coherent, structured and richer data, which will in turn make cross-border payments faster, more efficient and more reliable.

“In itself, ISO 20022 will not change the way payments will be processed. However, it will enable more straight-through processing of cross-border payments and ultimately enable a more frictionless and more instant cross-border payment infrastructure,” he told PYMNTS. 

But in order to reap the full benefits, widespread adoption will be key, Matthiessen noted, adding that banks will need to encourage market participants in the transition and provide a bridge to “pass on rich message format on the sequence of payment that exists in the cross-border world.” 

AI Requires Coherent, Rich Datasets 

The ISO 20022 migration is just one example of how the banking and finance industry is evolving. 

According to Matthiessen, the rise of artificial intelligence (AI) and the potential it has to transform and disrupt industries is another trend to watch, as many banks and FIs look to leverage the technology to improve the overall client experience, among other things.

For example, in cases where clients reach out to a bank’s call center with a service request, he said using AI and machine learning to integrate a speech-to-text feature can turn complex calls into readable text, minimizing friction from a back-end processing perspective. 

“This completely changes the way we can drive client experience and how we can also provide more digital automation of processes, particularly when you know the client or other market participants also use AI,” he said. 

In fact, the ability to connect information between two machines will completely change the way banks interact with clients in the future, he added, which will in turn reduce costs and make processes less prone to error.

Add-on services will be the second big application in the finance and banking space, Matthiessen said, particularly the ability to provide client-centric, data-driven solutions upon clients’ request. 

For this to be effective, however, having access to a coherent database on which AI and ML applications can run smoothly will be key, he said. 

Finally, he said AI can play a critical role in improving anomaly detection and fraud detection services, helping banks circumvent potential issues in the back-end system. “This will, in my view, not only help create a unique client value proposition, but also help improve and more consistently demonstrate the adherence to global and local regulations,” Matthiessen added.