Maybank saves $9.56m through small process changes

The bank says its initiatives have saved 2.17 million working days.

Malayan Banking Berhad (Maybank) has saved about $9.5m (MYR39m) through 1,500 small projects since 2024, as Malaysia’s biggest bank by assets gives teams more control over improving their processes.

“Sometimes, the larger savings come from simplifying the process around the model,” Siew Chan Cheong, group chief strategy and transformation officer at Maybank, told Asian Banking & Finance in an email interview arranged with Tech Week Singapore 2026. 

Maybank calls the approach “lite agile,” which lets teams fix productivity and workflow problems without waiting for approval from central technology staff or executives.

The initiatives have saved 2.17 million working days, according to Maybank’s 2025 annual report.
Maybank is also adopting a single leadership structure to reduce decision-making across separate teams and improve its sales operations.

Speed is another measure. Some 86% of what Maybank describes as its initial working versions of products were released within three to six months between 2024 and 2025. “They show the importance of short delivery cycles and visible outcomes,” Cheong said.

Maybank is applying the same approach to artificial intelligence (AI), payments, and its main banking systems.

Cheong said building a separate connection between each service and the bank’s main systems would quickly increase complexity. Maybank has instead simplified those connections whilst upgrading its main systems and improving its cloud computing, data management, and recovery processes.

The bank has integrated its corporate banking platform across Malaysia and Singapore, letting businesses see their cash positions in real time and use cash management services.

Maybank2E EzyApply also lets businesses apply online for corporate banking, trade finance, and cash management services with same-day approvals, Cheong said.

He said banks should judge technology investments by whether customers use them and whether they improve services rather than how quickly a particular technology is adopted.

He rejected the idea that banks could make their technology “future-proof” because payment standards, regulations, cyberthreats, and AI models will continue to change. “The better goal is to make change less disruptive,” Cheong said.

Banks should also track the cost of transactions, storage, moving data, and AI models to avoid overspending, he said.

“Cloud and AI are easy to consume,” Cheong said. “Without visibility of transaction, storage, data-movement and model costs, they are also easy to overspend on.”

Cheong said AI controls should depend on how much harm a particular use could cause. An employee using AI to draft a note requires different controls from AI used for credit decisions or direct customer communication.

Higher-risk applications require stricter testing, clear human responsibility, and closer monitoring, he said.

An investment ultimately needs to help the bank resolve cases faster, prevent losses, improve decisions, or serve customers better, Cheong said. “License fees and model costs matter, but they are only part of that equation.”