SIAgentNews.com
← All news
DevelopingAgents

Maybank delays customer-facing AI agents to prioritize auditability

Malaysia's largest bank is focusing on internal AI tools and system resilience while waiting for autonomous agents to become predictable enough for public use.

· First reported by Fortune

$2.5 billionMaybank's 5-year investment in tech and AI

The short version

  • Maybank is investing 10 billion Malaysian ringgit ($2.5 billion) over five years on tech, data, and AI under its ROAR30 plan.
  • The bank is currently using AI internally for staff recommendation engines and assistant tools.
  • Management will not deploy autonomous agents to customers until they are fully auditable and provide consistent answers.
  • The bank is using AI from ServiceNow to identify early system wear-and-tear in its Resilience Operations Centre.
  • Cyberattacks in Asia are becoming more dynamic and are suspected to be powered by AI.

Why it matters

For the average consumer, this means your banking experience won't be immediately overhauled by autonomous agents, but the underlying systems may become more stable. It highlights a growing trend where large financial institutions prioritize 'trust' and 'predictability' over being first to market with experimental tech.

Prioritizing trust over speed

Maybank, the largest bank in Malaysia, is taking a measured approach to the integration of agentic AI. While some of its competitors in Singapore have already begun allowing AI agents to handle tasks and pull transaction data for clients, Maybank is choosing to wait. The bank's leadership argues that the primary goal in banking is maintaining trust with the public.

Giorgio Migliarina, the bank's chief technology officer, noted that many people overestimate the immediate impact of new technologies while underestimating their long-term effects. He believes that rushing to deploy autonomous systems could jeopardize the bank's reputation. According to Fortune, Migliarina stated that the bank does not want to risk trust just to achieve a brief increase in productivity.

The bank's strategy focuses on ensuring that any AI used in a public-facing capacity is both predictable and auditable. This means the bank wants to ensure that if a customer asks the same question twice, the AI provides the exact same answer. Currently, the bank has various proof of concepts, but it is not ready to move them into a live production environment for customers.

Internal tools and infrastructure

While the bank is holding back on customer-facing agents, it is already heavily utilizing AI for its internal operations. These tools are designed to assist staff members by providing recommendation engines and digital assistants. By focusing on internal use first, the bank can refine the technology in a controlled environment before considering a wider release.

One specific application of this technology is the bank's Resilience Operations Centre, which is responsible for keeping the bank's systems running across Malaysia, Singapore, and Indonesia. The bank is equipping this hub with AI tools from ServiceNow to monitor the health of its infrastructure. These tools are designed to identify small degradations in the system, which Migliarina compares to the flickering of a lightbulb.

By identifying these early symptoms of wear-and-tear, the AI can prompt employees to take action before a major system failure occurs. This allows the bank to filter through normal activity to find the specific small issues that require human intervention. This proactive maintenance helps ensure that the core banking services remain available and stable for all users.

The venture capital mindset

Migliarina suggests that executives in heavily regulated industries like banking and telecommunications should approach AI with a venture capital mindset. In this framework, a company invests in many different projects or 'flowers' to see which ones will yield the best results. The goal is to identify the specific pilots that provide incredible returns and then focus resources on those successful areas.

This approach allows the bank to experiment with various AI applications without committing every resource to a single, potentially failing project. By letting many different pilots bloom, the bank can observe which ones perform well and which ones require more fine-tuning. This strategy helps manage the risks associated with new technology while still pursuing innovation.

The bank's overarching plan, known as ROAR30, involves spending 10 billion Malaysian ringgit over the next five years on technology, data, and AI. This significant investment provides the capital necessary to run multiple experiments simultaneously. The bank aims to find the most effective ways to integrate AI into its regional operations across the ASEAN economic power.

Security and regional growth

The rise of AI also presents new challenges for the bank, particularly regarding cybersecurity. Migliarina noted that cyberattacks in Asia have become increasingly dynamic and are suspected to be powered by AI. These attacks are changing in nature, making it more difficult for traditional security measures to keep up with evolving threats.

The bank also faces risks from the expanding network of fintech and payments partners, as every new connection creates a potential entry point for hackers. This complexity requires a sophisticated approach to security that can handle both the internal risks of AI and the external risks of a connected financial ecosystem. The bank must balance the benefits of these partnerships with the need to protect customer data.

Despite these hurdles, the bank remains committed to using technology to facilitate regional growth. Migliarina views the bank's role as a facilitator for the growth of ASEAN as an economic power. The ultimate goal is to achieve gradual but meaningful technological change that strengthens the bank's position without breaking the systems it relies on.

What to watch

  • How other regional banks respond to Maybank's 'auditable' requirement for AI.
  • The specific types of AI-powered cyberattacks hitting Asian financial institutions.
  • The timeline for Maybank's ROAR30 plan and its specific AI milestones.

Try it

Look for 'Resilience Operations' updates in your bank's annual report to see how they are using AI to prevent system outages.

Sources and links

First reported by Fortune. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news