Technology is beginning to alter the economics of banking in the Philippines in ways that extend beyond digitising existing products and distribution channels.

The more significant change is occurring in areas where conventional banking models have struggled with the cost of information, customer acquisition, risk assessment and service delivery. Payment data is giving lenders new ways to assess informal businesses. Artificial intelligence (AI) is allowing banks to extend elements of personalised wealth management to a broader customer base. Tokenised money is creating potential efficiencies in cross-border settlement and liquidity management.

These applications differ significantly, but they share a common economic proposition: technology can lower the cost of serving customers or executing transactions that were previously difficult to serve profitably.

That distinction matters in the Philippines. The country's digital financial infrastructure has expanded rapidly, but large parts of the economy remain informal and many banking relationships remain constrained by the economics of conventional distribution and underwriting.

The next stage of digital banking is therefore less about whether technology can improve the customer experience and more about whether it can change the boundaries of what banks can serve profitably.

Payment data is expanding the addressable market for SME lending

Small-business lending illustrates how data can alter the economics of risk assessment.

A sari-sari store (a small neighbourhood shop) may have regular customers, steady sales and a clear need for working capital while remaining difficult for a conventional bank to underwrite. Many small businesses operate informally, rely on cash and do not maintain financial statements or a clear separation between personal and business finances.

The result is an information gap rather than necessarily a shortage of economic activity.

Angelo Madrid, president of Maya Bank, said formal statistics for small and medium-sized enterprises (SMEs) understate the scale of entrepreneurial activity. While around one million SMEs may be formally registered, he estimated that the actual number could approach 20 million when informal businesses are included.

“The SMEs are more informal than you can imagine,” Madrid said.

For lenders, this creates a fundamental underwriting problem: whether the available information describes the individual borrower or the underlying business.

“The first question is that: Am I lending to the entrepreneur or to the business?” Madrid said.

Payment infrastructure can provide part of the missing information.

Maya has invested in acquiring SME merchants and building payment infrastructure across physical terminals, online activity and payment gateways. These transactions create a record of commercial activity that may otherwise be difficult for a lender to observe.

“That becomes a basis for underwriting them,” Madrid said.

The value of payment data is therefore not simply its volume, but the proximity to the economic activity. Transaction frequency, volume and patterns can provide signals about sales and cash flow that conventional credit files may not capture.

This is particularly relevant to nano and micro businesses, where the distinction between consumer and business finance is often blurred.

Moritz Gastl, president and general manager of Tala, said between 27% and 30% of the company's more than five million Philippine customers are nano entrepreneurs. Tala has disbursed about $3 billion in the country since entering the market in 2017, he said.

“These are your typical sari-sari stores, for example, that are using Tala loans to access working capital to be able to buy their goods and then distribute that to their customers,” Gastl said.

Tala combines customer, device and behavioural data to identify patterns associated with creditworthiness. This can help lenders identify customers who initially appear to be consumer borrowers but are actually using credit for business activity.

The commercial implication is significant. More granular data can influence not only whether a customer receives credit but also how the lender structures the product and manages the relationship.

The opportunity is consequently to expand the addressable lending market without relying entirely on greater formalisation first.

The competitive advantage is moving towards data aggregation and underwriting capability

Alternative data does not replace conventional credit information. Its value lies in combining multiple signals.

Gastl said payment and cash-flow data remain important but can be supplemented by behavioural, device and telecommunications data.

“Cash flow data and payments data hasn’t gone away. That is still incredibly important,” he said. “But paired with now the behavioural data and device data and telco data and all these alternative data sources that we have at our disposal, we are in a very unique position to target those nano entrepreneurs more so than the existing financial institutions.”

This points to a broader change in competitive advantage.

Historically, banks' ability to lend was closely tied to balance-sheet capacity, branch distribution, credit processes and access to formal financial information. Increasingly, the ability to aggregate and interpret data may become an additional source of advantage.

The economics are particularly relevant for smaller loans. Conventional underwriting can involve documentation, manual assessment and collateral requirements that make small-ticket lending costly relative to expected revenue.

Digital underwriting can reduce some of these costs by collecting information continuously, automating assessment and monitoring borrower behaviour after origination.

It does not eliminate risk. Rather, it potentially reduces the cost of identifying and managing it.

That distinction is important because the commercial opportunity in SME finance depends not simply on finding more borrowers but on making smaller and previously opaque borrowers economically viable to serve.

The constraint, however, is that data cannot resolve the structural incentives surrounding informality.

Gastl noted that consumer data in the Philippines is considerably more developed than corporate data, while small businesses may have limited incentives to formalise if registration brings additional tax and administrative obligations.

“The path to digitising the core economy, it’s not a digital bank problem; it’s an infrastructure problem,” Madrid said.

For banks and fintechs, the implication is that data-driven underwriting can expand access but will not eliminate the need for improvements in business registration, digital infrastructure and formal financial records.

AI is changing the cost equation for mass-affluent wealth management

The same economic logic applies to wealth management, although the problem is different.

The mass-affluent segment occupies a difficult position between conventional retail banking and private banking. Customers may have sufficient assets and increasingly complex financial needs, but not enough value to justify the full cost of traditional relationship-led wealth management. This creates a scale problem.

Jerry Ngo, CEO of EastWest Bank, sees the opportunity in the context of a growing population moving towards more sophisticated financial needs, including investment, retirement planning and wealth preservation. But wealth cannot be effectively segmented by assets alone.

“What we find more interesting is to be able to see what transitions people are undertaking, what goals in life that they're doing,” Ngo said. This is where data and AI can alter the service model.

Gauraw Srivastava, EVP and head of wealth management at UnionBank, said transaction patterns and behavioural signals can help banks create customer personas and identify where clients are in their financial journeys.

“Data is helping it so that you can read through certain patterns and you can put them into certain personas, you can help them achieve their goals,” he said.

The immediate opportunity for AI is therefore not necessarily automated investment advice. It is more efficient segmentation and allocation of human resources.

A customer who can complete a digital investment journey independently does not require the same level of intervention as one making a complex portfolio decision. AI can help determine where human attention is most valuable.

This potentially changes the unit economics of relationship management.

Rather than assigning a fixed service model to a wealth segment, banks can vary the level of human involvement according to customer needs and complexity.

“One-on-one doesn't mean that one RM can only manage 100 clients,” Srivastava said. “You can do hybrid.”

The objective is therefore not simply to reduce the role of relationship managers (RMs). It is to increase the number of customers each adviser can serve without reducing the quality of interactions that require judgement.

The economics of personalisation still depend on trust

Technology alone does not resolve the central challenge of wealth management: customers still need confidence in the institution and its advice.

Ngo argued that human interaction becomes more important as customers accumulate wealth and financial decisions become more consequential.

“Banking is about trust, it's about human connection,” he said.
EastWest's approach reflects a hybrid model in which technology supports the adviser rather than replacing the relationship.

“The biggest value proposition is to give you an RM,” Ngo said. “The whole point is you want to be served properly. It means you're valued, you have arrived.”

The distinction is commercially relevant. If AI is used to automate routine processes, analyse customer data and prepare advisers, banks can potentially deploy RMs across a larger customer base.

But if technology is used to remove high-value human interactions altogether, the bank risks weakening the trust and differentiation that justify a wealth relationship in the first place.

This makes the economics of AI more nuanced than simple cost reduction. The value comes from reallocating human expertise towards interactions where it produces the greatest commercial and advisory value.

There is also a longer-term economic consideration. “The same client who is perhaps worth a million pesos today may become a five million peso and 50-million-peso client over a period,” Srivastava said. “That's a long game.”

Tokenisation is targeting infrastructure costs rather than retail payments

The potential economic impact of tokenised money is clearest in a different part of the banking value chain: cross-border finance.

Domestic payments in the Philippines are already highly digitised. Consumers can transfer funds through mobile wallets and digital banking platforms with relative ease. The case for introducing another form of digital money at the retail level is therefore less compelling. Cross-border payments present a different cost structure.

Transactions can involve correspondent banks, foreign exchange processes, settlement windows and compliance requirements across jurisdictions. For large-value transactions, delays also have a liquidity cost. This is where stablecoins can provide a more direct economic proposition.

Nichel Gaba, founder and CEO of Philippine Digital Asset Exchange, said stablecoins can facilitate continuous cross-border movement of value while reducing reliance on some traditional intermediaries.

Dennis Bancod, assistant governor at the Bangko Sentral ng Pilipinas (BSP), said the economics become more material at institutional scale.

“When we're talking about hundreds of millions of dollars and in times of one to two days, the savings of it being a real-time transfer becomes much more glaring,” Bancod said.

The Philippines' remittance market offers an early testing ground. Bancod said that today, around $200 million to $300 million of remittances are already being channelled through stablecoins by institutional participants including major remittance companies and banks worldwide.

The significance is that consumers do not necessarily need to change their behaviour.

A bank or remittance provider can use stablecoins as settlement infrastructure while customers continue to receive conventional fiat currency. The technology therefore changes the provider's economics without necessarily changing the customer's interface.

That distinction is likely to be important in the development of tokenised finance.

The next opportunity is connecting tokenised money with tokenised assets

Stablecoins are only one component of a broader tokenisation architecture.

Tokenised deposits could allow commercial banks to represent deposits on tokenised infrastructure while retaining the underlying banking relationship. Wholesale central bank digital currencies (wCBDCs), meanwhile, could provide tokenised central bank money for regulated financial institutions.

The potential economic benefit increases if tokenised money can interact with tokenised financial assets on compatible infrastructure.

Payment, delivery and settlement could become more closely integrated, potentially reducing reconciliation, processing and intermediary costs.

In July, the BSP released its report on Project Agila, its pilot for a wCBDC. The proof-of-concept demonstrated the technical feasibility of using distributed ledger technology (DLT) and tokenisation to support 24/7 real-time interbank fund transfers and liquidity management. The findings guide the BSP’s wCBDC Roadmap, which will set out the central bank’s strategic direction for exploring and developing high-potential wCBDC use cases.

But the business case remains conditional. Banks and financial institutions would still need to invest in infrastructure, governance, controls and interoperability. Tokenisation creates economic value only if the savings and new capabilities exceed the cost of implementing and operating the new architecture.

There are also risks that technology cannot resolve.

Gaba said the residual risk associated with stablecoins remains at the issuer level. Blockchain technology can record ownership and settlement but cannot independently guarantee the quality of the reserves backing a token. That makes regulation and institutional trust central to adoption.

Bancod said the BSP's approach is intended to focus on outcomes rather than the underlying technology.

“The standard of the BSP will be about making sure that we are neutral,” he said. “We need to understand why it's there. We're more about the outcomes that come up, and then regulate those outcomes rather than looking at the technology and regulating that.”

Technology is changing the boundaries of the banking model

The three cases point to a common shift in Philippine banking.

Payment data can make informal businesses more assessable. AI can make personalised wealth management more scalable. Tokenised money can reduce friction in institutional settlement.

In each case, the technology has value because it changes an underlying cost structure. For SME lending, it reduces the information disadvantage that makes small borrowers expensive to assess.

For wealth management, it reduces the amount of human time required to understand and service a broader customer base. For cross-border finance, it has the potential to reduce settlement and liquidity costs.

This suggests that the competitive impact of technology will not necessarily be determined by which institution has the most sophisticated application or the largest digital customer base. It may increasingly depend on which institutions can integrate data, technology, capital and distribution into a commercially sustainable operating model.

That is also likely to reshape the relationship between banks and fintechs.

The Maya-Tala partnership provided one example. In April 2024, the companies launched a $48.5 million loan-channelling partnership under which Maya provided capital that Tala channelled through its mobile lending platform and credit-scoring capabilities.

The model combined different institutional strengths: bank capital and infrastructure with fintech distribution and specialised underwriting.

Madrid described this as part of a broader shift towards collaboration and open finance.
“I think the model for banking relationships is also kind of changing as well. You can look at our partnership as an application of open finance.”

Such models could become more important as financial institutions compete not only for customers but for access to the data and capabilities needed to serve them efficiently.

Technology expands profitable banking

The Philippines has already demonstrated that digital channels can change customer behaviour. The next question is whether technology can change the economics of the institutions serving those customers.

In SME finance, data is allowing lenders to assess economic activity that conventional credit models often miss. In wealth management, AI is allowing banks to segment customers and deploy human expertise more selectively. In cross-border finance, tokenised money is being tested where settlement and liquidity costs create the strongest economic incentive.

None of these developments makes traditional banking infrastructure obsolete, nor does technology eliminate the constraints imposed by regulation, trust, data quality or informality. Instead, the emerging model is one in which banks increasingly use technology to determine where conventional banking economics break down and whether those constraints can be reduced sufficiently to make new relationships viable.

The question is no longer whether a bank can put a product online. It is whether data can make a previously opaque borrower assessable, whether AI can make a previously uneconomic service scalable and whether tokenisation can make a previously costly transaction more efficient.