Small and medium enterprises sit at the core of Asia Pacific’s economic engine, yet they remain among the most underserved segments in formal finance.

Across the region, SMEs account for more than 97% of all businesses and contribute up to 53% of national GDP in some markets, according to the Asian Development Bank’s 2024 SME Monitor.

But why did we mention that SMEs are among the most overlooked sectors in formal finance in the same sentence when they are the driver for the region’s economy?

Well, despite their economic importance, access to credit remains constrained.

Mastercard estimated in 2024 that the SME credit gap in the Asia Pacific stands at US$2.5 trillion, a figure echoed by the International Finance Corporation’s 2025 MSME Finance Gap Report, which warns that demand for financing continues to outpace supply.

That’s a couple of years ago. The number these days could be very different.

Traditional banking models have now long struggled to serve this segment at scale.

Legacy technology stacks, fragmented regulatory frameworks, and manual workflows continue to slow credit decisions, even at a time when fintech players redefine expectations around speed, automation, and accessibility.

Joe Udomdejwatana, Business Development Director for Asia Pacific at Axe Finance, believes banks remain stuck in incremental digital transformation, and they have a lot to do.

Speaking during the Singapore FinTech Festival, he explained why he believes, Lending 3.0 represents a fundamental shift in how financial institutions can modernise credit decisioning and reach the region’s underserved enterprises.

The Illusion of Digital Transformation in SME Lending

Banks often highlight digitalisation initiatives as evidence of progress, but Joe argues otherwise.

He said,

“What banks do at the moment is actually transforming digital from [a] manual process.”

Joe pointed out the example like how relationship managers often still gather SME documents manually before entering them into CRM systems, with analysts relying heavily on Excel for financial spreading.

The underlying issue, he explained, is that digital tools have been layered onto legacy workflows rather than replacing them with streamlined processes.

Many banks continue to stitch together multiple decisioning engines, data warehouses, and third-party systems in an attempt to automate lending, but the resulting complexity often slows down decision-making instead of accelerating it.

Fragmented technology stacks can create operational bottlenecks that make it difficult for banks to scale SME lending efficiently, even when demand is strong and data is readily available.

Why Legacy IT Budgets Keep Banks Stuck

Plus, the constraints of fragmented, legacy infrastructure are not only just technical but also financial.

Forrester estimates that traditional banks spend around 70% to 80% of their IT budgets just on maintaining existing systems alone. Which tends to mean that they often leave limited capacity for innovation.

Composable architecture is often discussed as a solution for that, and the Business Development Director for Asia Pacific at Axe Finance agrees.

But he did put an emphasis that modernisation must occur without disrupting core banking operations.

And what does he mean by that? Joe gave a perfect analogy.

“Think of fully composable as something … like you’re renovating a room in a house, [that] you’re still living in it, and you don’t want it to disrupt your lifestyle while [the renovation is taking place],” he said.

Rather than replacing core systems outright, banks can add modular capabilities on top of their existing infrastructure.

The layered approach enables institutions to roll out new capabilities while preserving system stability, a non-negotiable requirement in heavily regulated financial services environments.

The Black Box Problem in AI-Driven Credit

Artificial intelligence has become central to a lot of segment, and modern credit decisioning is one of it. However, it lacks explainability which remains a major concern for regulators and risk leaders.

Opaque AI models create accountability challenges, particularly when decisions must be justified to regulators and internal risk teams.

“Regulators will not accept [when you tell them] that you approve [a loan] because of AI,” Joe smiled while he gave this example.

To address such issue, he advocates hybrid models that combine AI with traditional statistical approaches.

@fintechnewsnetwork

Banks Can No Longer Hide Behind AI Regulators will never accept “because the AI said so” for credit decisions. So, risk officers are now required to have a transparent and auditable trail to ensure compliance. Is your institution prepared for the era of explainable AI? #Fintech#AI#Banking#RegTech#AxeFinance

♬ original sound – Fintech News Network – Fintech News Network

Rather than relying entirely on machine learning, banks should also ensure decisions remain traceable and auditable, with clear explanations for each approvals or declines.

Why?

Because explainability has increasingly becoming a regulatory requirement rather than just an optional feature, as automated decisioning becomes more widespread across financial services.

Regulation as a Structural Barrier Across Asia

Regulatory fragmentation adds another layer of complexity for banks operating across Asia Pacific.

“In this region, I think the reason is because of [the] diversity,” Joe mentioned.

What he means by that is not only because of the divergent nature of the region, but also because each market is governed by its own central bank and regulatory framework, creating significant variations in workflows, credit scoring requirements, and compliance processes.

Hence, coordinating compliance across multiple jurisdictions adds operational strain for regional banks, particularly when regulatory changes must be implemented quickly.

“One challenge is … how do we actually centralise all that at the region … to allow changes [to be made] real time and rapidly,” he added.

But Joe has an answer to that.

He admits that platforms that support multi-entity frameworks could help banks manage regulatory requirements centrally.

Which all can be done while still complying with local mandates, reducing the operational burden created by fragmented compliance processes.

Empowering Business Teams With No-Code Tools

Beyond these issues though, Joe Udomdejwatana sees a growing shift towards empowering business users such as compliance officers, risk managers, and product teams to configure systems directly.

He believes that no-code tools will allow policy updates, workflow changes, and regulatory simulations without any form of intervention from the developer.

Not only that, decentralising configuration capabilities can significantly shorten time to market and allow banks to respond more rapidly to regulatory changes.

Joe said that giving business teams direct control over rules and regulatory simulations helps institutions reduce dependence on long development cycles while increasing agility in lending operations.

Embedded Lending as a Massive Opportunity

Embedded finance also represents a major growth opportunity, particularly within B2B ecosystems.

In Asia Pacific alone, embedded lending is estimated to represent a US$78 billion market.

Speed however, will be the defining factor.

“Embedded financing lives and die basically on speed of decisioning in seconds, not days,” said Joe.

Traditional banks often struggle to match fintech agility, where loan approvals can happen in seconds rather than days.

He strongly believes that composable architecture and AI-driven models can help banks compete by enabling faster decisioning and seamless integration into digital platforms.

Joe is also on the side that technology can give banks the capabilities to compete with fintech players.

However, success will also depend on how well banks prepare their operations and integrate into digital ecosystems.

Lending 3.0 and the Need for Intelligent Adaptability

For what lies for the future, Joe describes the next phase of credit as Lending 3.0, driven by real-time data and adaptive decisioning systems.

“I would say the term is intelligent adaptability,” he indicates.

Modern lending platforms must ingest, interpret, and act on data continuously rather than relying on periodic assessments.

Banks must feed real-time performance data, behavioural signals, and application information directly into their risk models to maintain accuracy.

Such transition will now help to redefine lending from periodic underwriting assessments to continuous risk monitoring, requiring platforms capable of sensing changes, interpreting data, and acting in real time.

Flexibility and agility will become essential as long as lending models evolve towards real-time intelligence.

From Legacy Maintenance to Intelligent Banking

Ending his thoughts, Joe believes that banks must move beyond maintaining legacy systems and focus more on becoming intelligent service providers.

All of the things he mentioned beforehand, the composable platforms, explainable AI, embedded finance, and real-time adaptability, all represent a roadmap for banks seeking to close the SME credit gap and compete with fintechs.

Technology alone, however, will not solve the structural challenges.

Banks must be able to rethink workflows, regulatory coordination, and decisioning models in order to truly unlock Asia’s vast SME potential.

Joe Udomdejwatana highlighted that the region’s credit gap remains one of the largest untapped economic opportunities.

Whether traditional banks can evolve quickly enough will determine who captures it in the next decade.

Watch the video as Joe and I unpack the limitations of basic digitisation and explore what true Lending 3.0 strategies could mean for banks across Asia Pacific.