More than 70% of Southeast Asian adults remain unbanked or underbanked. That is not a development statistic. It is the largest untapped credit market on earth.
Millions of these consumers are not poor credit risks. They are simply invisible to the underwriting models banks have relied on for decades.
Southeast Asia’s unbanked population represents a $1.5 trillion financial services opportunity.
Digital financial services in the region’s six largest economies could generate up to $60 billion annually, but are currently on track for $38 billion.
That $22 billion gap is the commercial cost of financial exclusion. The challenge: expand credit responsibly without increasing risk or cost-to-serve.
“Across Asia-Pacific, financial inclusion remains a significant opportunity — over 1 billion people in the region still lack access to formal financial services.”
said Aashish Sharma, Head of Digital Strategy & Innovation, FICO.
The Challenges
- Credit Risk With No Credit History
Fitch noted that banks in the Philippines, India and Vietnam carry the highest risk appetite in APAC, partly because expanding financial inclusion means moving down the credit curve.
The consequences are visible: Vietnam’s NPL ratio has climbed to 5.4%, Cambodia’s to 7.0%, and the Philippines’ to 3.44%, a nine-month high as of May 2026.
Large segments of this population are thin-file or credit-invisible. No salary slips. No credit bureau record.
No formal collateral. Traditional underwriting cannot assess them, not because they are uncreditworthy, but because the data does not exist in forms banks know how to read.
- The Weight of Geography
As of March 2025, 468 cities and municipalities in the Philippines had no bank at all, 45% of them in Mindanao.
One-third of all Philippine branches sit in Metro Manila. Digital banking became the only viable inclusion strategy, but digital has limits.
Financial literacy among low-income Filipino users sits at just 34%. Having a smartphone is not the same as being financially included.
Thin physical presence also means banks extend credit with less ground-truth context about borrowers, a structural driver of higher NPL floors for digital-first lenders.
GCash’s lending arm runs at 4.9% NPL, Maya Bank at 5.77%, against 1.8% at BDO.
- Regulatory Balancing Act
Regulators across APAC want broader inclusion and prudent lending, and tighter fraud controls, simultaneously.
Regulatory controls are fragmented across APAC but are built around key priorities such as licensing, tiered products, consumer protection, and enforcing real-time operational resilience.
These goals are in structural tension. Every design decision that lowers the barrier for a legitimate unbanked borrower also increases the NPL risks and lowers it for a fraudster.
The Philippines’ BSP has been among the most active: its digital payments push smashed its own targets two years early, and its June 2025 AI model risk management guidelines require validation frameworks for all supervised institutions.
But a new 2026 rule requiring rural digital banks to keep 70% of customers within their physical operating area risks freezing digital banking’s reach into underserved communities, while also signalling a need for a wider ecosystem.
The AI-Led Way Forward
Alternative Data Scoring: Seeing the Full Borrower
AI-facilitated credit scoring expands beyond the bureau, reading patterns in spending behaviour, payment velocity, account balance trends, cash flow dynamics, and device metadata.
The result: a richer, more accurate picture of creditworthiness than any credit file provides.
At origination, this approach enables lenders to deploy rules-driven strategies that expand access while maintaining strong risk controls.
These strategies can be rapidly adjusted as market conditions shift or to meet the segment’s needs.
A truly intelligent decisioning platform takes this a step further – the ability to operationalize machine learning directly within origination workflows.
Every decision is governed by transparent, auditable logic, giving business and risk teams full visibility into what drove an outcome and why.
Grab Finance offers one example of this approach in practice.
It deployed 22 AI decision workflows across six Southeast Asian countries in under eight months, lifting credit offer eligibility by about 50% for previously ineligible users, according to FICO.
The Stakes
The banks that get AI-led inclusion right will not only grow faster in the world’s highest-growth region, but will also own a generation of first-time borrowers and loyalists with decades of financial life ahead of them.
The banks that get it wrong will build the next NPL crisis.
The $22 billion gap between where digital financial services are and where they could be is not a technology problem.
It is a decisioning problem. Applied intelligence, deployed with rigour, transparency, and genuine consumer protection, is how it gets closed.
Discover the success of GXS, ANZ New Zealand Banking Group, and AU Small Finance in creating a continuous inclusive lending intelligence loop from first credit assessment through to ongoing customer engagement.
Download FICO’s whitepaper, Opening Credit Opportunity Without Opening the Door to Risk here.
Featured image: Edited by Fintech News Singapore based on an image by FICO via gettyimages.
