Ashish Sharma has seen AI fads come and go, but what he’s seeing now in Malaysia’s banking scene feels different. Not because the technology itself has reached some mythical maturity, but because the questions are starting to shift.

“Banks aren’t just asking what AI can do,” he observes. “They’re asking what it should do, and how to make it stick.”

As Country Sales Director for Malaysia at Backbase, Ashish works closely with traditional banks seeking to modernise in a fast-changing digital environment.

He sees signs of maturity as banks move beyond the hype cycle — but most are still stuck in the pilot phase.

According to a 2023 Bank Negara Malaysia (BNM) survey, over 80% of Malaysian banks have begun AI initiatives. Yet fewer than 15% have moved into full-scale production.

Meanwhile, MDEC reports that Malaysia’s national digital investments hit RM163.6 billion in 2024, with AI solution providers generating over RM1 billion in revenue between August 2023 and July 2024.

So while the ambition is clearly there, the real question is, who’s actually making it work in practice?

Where Malaysian Banks Really Stand on AI 

Some banks are charging ahead.

For starters, the digital-only Ryt Bank is preparing to launch a full-scale AI-powered platform in a field where most incumbents are still catching up in a loop of pilots and proofs-of-concept. Even so, real progress is happening across traditional players in Malaysia.

We can see how CIMB is applying AI to speed up SME onboarding. Launched back in 2016, CIMB’s EVA chatbot (which was later updated in 2018 to include Artificial Intelligence (AI) and Natural Language Processing (NLP)) was developed to better support SME customers’ banking needs in a fast and secure manner.

The bank said the platform is available 24 hours a day, seven days a week and can handle a large number of simultaneous queries from SMEs.

Another example came just last year, where Hong Leong is reported to have also integrated AI into its collection and telemarketing operations, which reported a 15-fold increase in productivity and an 86% reduction in related costs.

According to the bank’s CEO, Kevin Lam, the system uses AI voice bots and has taken over the workload of more than 20 human agents. The current application is expected for outbound calls use, but there were also plans to expand the AI’s capabilities to manage more complex customer service interactions.

“Banks have all the data, but it’s scattered,” Ashish answered. “That’s what’s holding back serious AI capabilities.”

“That’s where a platform-first strategy pays off,” he explains. “When everything is stitched together, including customer & employee journeys, channels, and data, the bank can actually start delivering value from day one.”

By connecting disjointed data systems and embedding intelligence across the customer lifecycle, banks can reduce friction, boost service efficiency and deliver personalised experiences that are compliant, consistent and contextual.

Why AI Remains Fragmented

Ashish sees the root of the challenge: disconnected systems.

“Most banks have siloed infrastructures across onboarding, mobile banking, loan origination — and those silos result in fragmented data.”

While some banks try layering these “fixes” on top of their legacy systems, these bank-aid fixes rarely scale. What’s needed instead is a composable architecture: modular, progressive and pragmatic.

“You can begin with AI that validates docs (salary slips, IDs), then build from there,” he explains.

This modular approach is core to Backbase’s platform strategy. Rather than patching tools on top of legacy infrastructure, the platform introduces a unified engagement layer that connects front-end experiences with back-end systems.

This enables banks to progressively embed intelligence into critical customer journeys without the overhead of rebuilding from scratch.

As banks move beyond pilots, the opportunity lies in embedded intelligence, where AI is not just reactive but can help trigger actions across journeys.

With a platform-centric approach, financial institutions gain the ability to automate routine tasks, reduce friction and improve speed to serve, all while staying in full control of compliance and experience.

Tech Alone Won’t Get You There

For many banks, the bigger challenge isn’t the technology, it’s the people. Like most countries in the region, Malaysia still faces an AI talent gap. Ashish explains how Backbase works closely with its customers to address that.

Backbase’s solution: an “adopt and build” strategy.

“We co-implement the minimum viable product with the bank’s tech teams. Then we also train and certify them. Once enabled, they take over future enhancements. That’s how it becomes sustainable.”

This approach accelerates go-lives, supports long-term innovation and improves retention. Engineers are more motivated when they’re building — not just maintaining.

Ashish also warns against AI projects run solely by tech teams.

“When tech and business align, you start seeing real business outcomes.”

For that to happen, banks need a culture shift. One that’s more open to testing, learning, and adjusting as they go.

AI Ethics Is a Moving Target and Banks Know It

As banks are moving faster with AI, regulation is trying to play catch-up. Malaysia introduced its AI Governance and Ethics Guidelines (AIGE) in 2024, a voluntary set of principles, less prescriptive than that of Singapore’s rules, but still a step forward.

The Personal Data Protection Act (PDPA), however, hasn’t yet caught up with automated decision-making. That puts the onus on banks to tread carefully.

“As banking is a highly regulated industry, we make sure all our AI use cases are designed within the guardrails,” Ashish says. “Every feature gets validated with our banking clients before it goes into our product launch. I believe that’s also why our Chief AI officer is a banker, to ensure checks and balances.”

Ashish sees regulation and innovation as partners in progress, not rivals. And he’s optimistic about AI’s inclusive potential, from local-language voice banking to safeguards for vulnerable users.

“Voice-based conversational banking can help those who aren’t comfortable typing in English. It’s small, but it matters.”

Forget the Flash. Focus On What’s Worth Building?

Ashish’s message to banks is simple: focus on what solves real problems.

“Are we doing this because it looks good, or because it solves something critical?”

Build what lasts. Startups can afford to experiment. Big banks, on the other hand, can’t. Every new feature brings cost, risk and complexity.

This is where banks must assess whether their AI efforts are orchestrated or opportunistic. Without a platform to unify data, workflows and intelligence, even the best algorithms won’t scale.

The future belongs to those who embed AI into the fabric of their customer journeys, not just their innovation labs.

From Pilot Projects to Scaled Impact

“To truly operationalise AI,” Ashish concludes, “banks must shift from project-based experiments to a unified platform strategy that delivers personalised, scalable value across all customer segments.”

Malaysia’s next wave of AI transformation won’t be built on isolated tools or quick wins. It will be driven by composable foundations — intelligent, secure, and human-centric.

With the right platform in place, banks can embed AI where it delivers the most value, then scale with confidence.

The result? Not hype. Just better banking.

For more information about how Backbase empowers Malaysian banks in AI operations, visit their website here.

Featured image: Edited by Fintech News Malaysia, based on images by EyeEm and gesrey via Freepik.