Asia-Pacific finance shifts from AI spending to measurable value

Asia-Pacific financial institutions are entering a new phase of artificial intelligence adoption, moving beyond pilot projects and technology spending towards measurable business value as customers demand faster, more personalised and always-on services.

Speaking at the Huawei Thailand Digital & AI Summit 2026, Martin He, Senior Solutions Architect in the Finance Account Department at Huawei Thailand Enterprise Business, argued that banks would gain little from running multiple AI pilots unless the technology produced clear operational and commercial results.

“The future will not belong to banks with the most AI pilots. It will belong to banks that turn AI into results,” he said.

AI drives the real-time economy

He described 2026 as a turning point, with AI evolving from a tool that waits for instructions into “productive intelligence” capable of planning, collaborating, executing tasks and learning alongside people. This shift, he argued, could reshape banking through deeper digitalisation and more intelligent decision-making.

At the centre of that change is what he called the “real-time economy”, the next stage after instant payments and instant retail.

In many Chinese cities, consumers can already order groceries, daily essentials or urgent medicines and receive them within 30 minutes. The significance lies not only in rapid delivery but also in the real-time links between customers, merchants, warehouses, logistics providers and financial services.

For banks, that model requires always-on transactions, real-time data processing and more agile, cloud-ready applications. Instant payments have changed how people pay, while instant retail is beginning to alter how the wider economy operates.

“The model may look different in every market, but the direction is the same: business is becoming more real-time,” He noted.

Banks seek measurable returns from AI

Agentic AI could take this further by helping financial institutions respond to customer needs as they emerge. He envisaged a future in which customers receive more personalised services while employees work with AI-powered partners to process information, make decisions and complete tasks more quickly.

Rather than being confined to isolated functions, AI is increasingly being applied across banks’ value chains, from internal productivity and operations to marketing, risk management and customer-facing services.

Figures presented at the summit showed that AI investment by Asia-Pacific financial institutions could rise about eightfold, from US$440 million in 2025 to US$3.5 billion by 2030.

The presentation also cited DBS Bank, where the reported economic value generated by data analytics, AI and machine learning rose from S$180 million in 2022 to about S$1 billion in 2025.

For He, the figures illustrated a broader transition across the region from AI spending to measurable AI value. However, he cautioned that investment alone did not guarantee returns, with many enterprise AI projects still struggling to produce demonstrable results.

The strongest gains, he argued, would come when AI was closely linked to business outcomes, including better customer experiences, faster operations, stronger risk controls and more informed decisions.

Skills remain a critical constraint

Yet the expansion of AI in finance also exposes a major constraint: a shortage of people with the skills needed to build, operate and govern the technology.

“AI cannot scale unless talent scales with it,” He stressed, adding that the lack of skilled professionals was one of the most common concerns raised by financial-sector customers.

Huawei plans to train more than 10,000 finance-industry professionals worldwide over the next three years through strategic workshops, practical training and professional certifications for executives, business teams, technical staff and AI practitioners.

He also underlined that technology alone was insufficient. Financial transformation requires innovation, strong engineering, an open partner ecosystem and effective local delivery, alongside employees who understand how to apply AI responsibly to real business problems.

The next stage of AI in finance, therefore, will not be measured by the number of models or pilot projects a bank launches. It will be judged by whether institutions can combine intelligent systems with skilled people to deliver better work, better decisions and better banking in an economy that increasingly operates in real time.

AI must solve real customer and business problems

Dr Theerachai Chanyaswad, Chief Data Officer at KASIKORN LINE, argued that the next phase of financial AI would be judged by whether it could turn data into faster decisions, broader access to credit and more responsive customer service.

LINE BK has long used alternative data to assess applicants with limited conventional credit histories. That experience, he noted, showed that data could create value only when it addressed a clearly defined customer or business need.

“AI is now involved in everyone’s lives, whether we want it or not,” he remarked. However, adoption should not begin with the technology itself. “The starting point is the pain — the customer’s pain and the business’s pain.”

He pointed to areas in which AI could move financial services closer to real-time operations, including reviewing customer interactions at scale, identifying service problems more quickly and enabling more personalised communication. 

Tasks that once relied on small samples or lengthy manual checks could increasingly be assessed across complete sets of customer interactions.

The goal was not simply to automate existing work but to help financial businesses respond faster while keeping operating costs under control.

“We want to grow the portfolio and the business without costs rising at the same pace,” he explained.

Insurance cannot be sold at yesterday’s speed

Sochipun Vajropala, Chief Executive Officer of DataOne Asia (Thailand), argued that the insurance industry was also moving towards real-time, personalised services as customers expected companies to understand their needs and respond immediately.

Traditional insurance models have focused on calculating and paying claims. The emerging model places greater emphasis on preventing risk, controlling costs and providing services built around individual circumstances rather than on selling a standard policy in isolation.

Younger customers, in particular, expect digital services to respond with little or no delay, she noted.

That shift is changing the meaning of an insurance product. Health cover, for example, is increasingly linked to preventive care, medical consultations, clinics and access to medicines. Other opportunities are emerging in retirement, investment and electric vehicles as customers seek services that fit their daily lives.

AI plays a central role in processing customer information, assessing risk and personalising services at the speed required by this new market. However, Sochipun stressed that human involvement would remain necessary for complex or high-risk decisions.

Fraud controls must keep pace with real-time finance

Keerawan Panyasen, Account Director at TrustDecision Thailand, focused on another consequence of the real-time economy: financial crime now moves at the same speed as legitimate transactions.

“Fraud is no longer only a security problem,” she warned. “It is a bottleneck affecting business, operations and compliance.”

Customers increasingly expect banks and their fraud-detection systems to identify suspicious behaviour even when the account holder has technically authorised a transfer. A failure to intervene can damage trust, reduce revenue and slow the broader shift towards digital financial services.

The challenge is that fraud no longer appears as a single unusual transaction. Criminal networks may operate across accounts, devices, sessions and channels, while older rule-based controls assess each event in isolation. 

This can leave investigators overwhelmed by false alerts and unable to stop genuine threats quickly enough.

“The real challenge is not detecting more fraud,” Keerawan argued. “It is detecting it faster, deciding more accurately and reducing friction for genuine customers.”

AI can help financial institutions analyse patterns throughout a customer’s digital journey and identify anomalies before money moves. In a real-time system, however, speed is decisive.

She also cautioned against removing human judgement from sensitive decisions. Suspicious or complex cases still require experienced investigators, clear governance and auditable records.