Malaysia's Finance Firms Hit a Wall: Legacy Systems, Not AI Shortage, Block Progress
Malaysian banks and insurers are pouring money into artificial intelligence, but their real problem sits in the basement: outdated infrastructure, messy data, and processes designed decades ago that no amount of new technology can fix.
Finance executives gathered at the Asian Banking & Finance x Insurance Asia Summit in Malaysia this week made clear that AI adoption isn't a technology problem. It's an operational one.
"We need more money. We need more support from the board, as well as our government," said Tok Puan Datin Ezreen Eliza, head of Transaction Banking - Securities Services at CIMB Bank and CEO of CIMB Trustee. But the barrier isn't cost alone. It's the work required before AI can actually help.
The Square Wheels Problem
David Brandl, Chief IT Officer at Allianz Malaysia, described the challenge bluntly. Operational teams are too consumed with daily work to step back and rethink their processes.
"It's like a person 200 years back pulling a cart with square wheels," Brandl said. "When offered a better solution, the person says, 'Sorry, I don't have time for your ideas.'"
Allianz Malaysia made simplification a strategic priority. Teams now review products, processes, and systems to decide what to change, what to keep, and what to eliminate. The goal isn't cost-cutting alone-it's better service delivery.
Regulators Complicate Everything
In Malaysia's financial sector, process redesign hits another snag: multiple regulators with overlapping rules. Ezreen Eliza pointed to custody and trustee services, where firms answer to both the Securities Commission and Bank Negara Malaysia.
"We have two sets of regulators," she said. Adding new technology to regulated workflows means redesigning processes while staying compliant with both authorities-a complexity that slows everything down.
AI Can't Replace Judgment-Yet
Vivien Tan, SVP and group head of SME Credit at Alliance Bank Malaysia, offered a direct assessment: "Can AI do everything for now? My answer is no."
AI works well with structured data. But SME banking relies on seasonal patterns and non-standard terminology that AI struggles with. Data quality remains the constraint. "Garbage in, garbage out still applies," Tan said.
Ezreen Eliza shared an internal CIMB discussion on whether AI could handle KYC and account opening with a three-person team. The answer: no. AI can flag exceptions and simplify workflows, but it can't remove people from the process entirely.
Amine Chiha Setté, head of IT at Generali Malaysia, emphasized that finance firms need to think beyond AI as a standalone tool. "We are not talking here about using AI basically to respond to a client, but we have to build the solution," he said.
Legacy Systems Are the Real Barrier
Tan was direct about what's holding Malaysia's banks back. "The legacy systems is actually problem. It's really painful to actually go through that kind of transformation and use that kind of money."
Even larger banks struggle with aging infrastructure. While AI tools are likely to become cheaper over time, banks first need to invest heavily in cleaning up their data and replacing old systems-a multi-year effort with uncertain ROI.
Rethinking ROI
Gerard Ang, country head of Financing & Securities Services at Standard Chartered Bank Malaysia, pushed back on the cost-cutting narrative. "A lot of people look at costs thinking that their revenues will continue to stay the same. What if you maintain your cost, increase your values?"
Some technology spending, including digital assets, should be viewed as long-term investment in future business models rather than near-term expense reduction.
The panel was generally positive on Malaysia's prospects, though executives acknowledged the work required. Technology should support human decision-making, not replace it. And before AI can do either, the foundation has to be fixed.
Learn more about AI for Finance or explore the AI Learning Path for CFOs.
Your membership also unlocks:
