What does the ‘AI-first’ bank actually look like in Southeast Asia?

This article summarizes an episode of Fintech Fireside Asia’s video series featuring top leaders from Malaysia’s banks.

Photo credit: Shutterstock

Leaders from Malaysia’s digital and incumbent banks are applying AI across every layer of operations. As they navigate new opportunities, their primary challenge is slow human adoption, not the technology itself. Their success will depend on managing this adoption gap while at the same time defending against AI-driven fraud.

AI moves beyond the back office

AI is not just for background tasks anymore. Banks are now using it in their work with customers, like in debt collection and sales.

Finding the best way to contact customers
Previously, debt collection relied on mass contact, wasting time on methods customers ignored. Banks now use AI to learn each customer’s preferred contact method, aiming to make the first attempt the most successful one.

William Streitberg, CTO of Hong Leong Bank, explains, “we’re now looking from a collections aspect the best contact points for customers. So how do they best respond? Is it through email? Is it through phone call? Is it through SMS? … We don’t want to waste cycle time and we don’t want to waste our people.”

Finding serious buyers
In sales, too much time is spent contacting people unlikely to buy. AI can now handle the initial contact to identify serious buyers.

Diana Boo, CMO of Boost, notes, “we just on Monday rolled out our conversational voice AI that is a sales assistant. So today if you apply for a loan, an agent AI will call you and to help you verify what are the things that you want … it helps our BD team in terms of funneling quality customers.”

The AI vs. AI fight

These new tools have also created new weaknesses. Banks are now facing fraud attacks made by AI. This situation creates a fight where banks must use their own AI to block the advanced tools criminals use.

The new type of digital criminal
The stereotype of a lone hacker is outdated. Today’s financial crime is organized and high-tech, with criminals using AI to create realistic fake information.

GXBank CEO Kaushik Chowdhury warns, “let us not underestimate that threat actors are using generative AI to attack you. I was really surprised myself in the last 6 months, when we really tried scaling our digital lending, as to how sophisticated some of these fraud actors are. They are not amateurs working on a laptop trying to do metadata changes on PDF.”

A fight in the code
Rule-based fraud detection systems cannot stop adaptive attacks. The only effective defense against an AI-powered threat is another AI.

Chowdhury continues, “They are using generative AI to do that, and we are using generative AI models for our IP… It’s AI versus AI, and the stage is fraud detection, deflection, and monitoring.”

Freeing up the front lines

As AI does more of the simple work, the goal is changing from replacing people to helping them. Banks are giving employees AI tools to help them work.

Letting everyone build with AI

Technology is moving faster than people

The problem with old habits