In an interview with The Digital Banker, Jackson Oh, Chief Technology Officer at ANEXT Bank, discusses how live transaction data, embedded lending and automated credit assessment could reduce friction in SME financing while maintaining the governance, security and resilience required of a regulated bank.
Singapore’s latest fintech funding push comes as financial institutions accelerate the adoption of artificial intelligence, digital infrastructure and other technologies across financial services. On 31 August, the Monetary Authority of Singapore committed SGD220 million (USD173 million) over three years under the renewed Financial Sector Technology and Innovation scheme, supporting areas including AI adoption, financial infrastructure and innovation.
For SME banking, one important application of technology sits in credit assessment. Smaller businesses often lack the financial history, documentation or established banking relationships available to larger companies, creating friction when they seek financing.
ANEXT Bank has taken a data-led approach. Its CreditNow facility uses alternative data and embedded lending across its partner network, reducing application time to as fast as one minute. Selected existing customers can access credit lines of up to SGD300,000 (USD 236,486), subject partly to at least six months of activity with the Bank.
For Oh, the opportunity involves more than digitising an existing process. The underlying credit problem needs a different approach. “The core decision was to stop treating the loan application as a document-gathering exercise and treat it as a data problem’, he explains.
That means shifting from historical paperwork towards live signals generated by the business. CreditNow draws on transaction data and automated scoring to assess an SME’s activity, potentially giving the lender a more continuous view of its financial behaviour.
From paperwork to operating data
Oh’s career across VMware, Samsung and Bank of America has centred on building complex infrastructure that remains largely invisible to users. At ANEXT, that experience informs his approach to SME banking, where newer businesses can face particular barriers to credit.
Traditional underwriting can require businesses to collect, verify and interpret multiple documents before a decision can proceed. CreditNow removes much of that friction by using transaction data, but the engineering work behind the customer-facing one-minute application remains substantial. “The headline number of ‘as fast as one minute’ is really the visible tip of a set of decisions made much further upstream,” Oh says.
ANEXT Bank needs reliable transaction feeds from partner platforms, while dealing with gaps in data quality, differences in formats and timing issues. Oh describes data access as the easier part of the process; ensuring that the information remains complete and usable presents the harder engineering challenge.
The approach also reflects Oh’s view that financial institutions can support SMEs beyond the point of providing credit. He recalls a South African hairdresser who built her business from home before formalising it and seeking financing to open a shop. She hired a consultant to prepare a business plan, recognising the need to establish a credible business profile.
Many smaller business owners lack the resources or foresight to take those steps themselves. Technology could help banks provide some of the support that sits around financing, rather than simply assessing an application at the end of the process.
“I believe banks, especially digital banks and fintechs, can contribute by being more than just lenders”, Oh explains.
Faster decisions, without lighter controls
Using alternative data does not remove the need for rigorous credit assessment. Instead, Oh argues, it can give lenders more evidence to work with.Transaction behaviour can provide a continuing view of how a business operates, while conventional financial statements offer a historical snapshot. For newer SMEs, that distinction can matter when a limited financial record does not fully capture current business activity. “The instinct people have is that alternative data means lower rigour. Whereas in actual fact it’s the opposite; it actually means more evidence, assessed more consistently.”
That model also needs to remain explainable. Automated scoring cannot become a black box, particularly within a regulated banking environment. “A model that risk and regulators can’t interrogate is a non-starter in a MAS-licensed bank’, as Oh points out.
ANEXT Bank’s credit modelling team continually validates its models using refreshed datasets, including point-of-sale transactions, banking activity and credit bureau information, according to Oh. The emphasis remains on combining automation with governance rather than replacing conventional risk controls.
Building for speed and scale
ANEXT Bank’s data-driven approach extends into embedded finance, connecting its infrastructure with third-party platforms such as EPOS systems. Bringing financing into the workflows where businesses already operate can reduce friction, but it also introduces dependencies outside the bank’s direct control.
The Bank maintains clear boundaries between its infrastructure and partner environments, with APIs governing data exchanges and controls protecting information in transit and at rest.“We never depend on the partner’s environment being trusted”, Oh says.
The same principle applies to engineering speed. Oh argues that compliance and resilience should form part of the development process rather than operate as barriers at the end of it. “Compliance and resilience aren’t the brakes on speed. When done properly, they’re what let you move fast safely.”
At ANEXT Bank, compliance checks and security scanning form part of the engineering pipeline, while risk and compliance teams remain close to product and engineering decisions.
Scale will place greater demands on those controls. Oh says the Bank plans to extend CreditNow to more than 100,000 merchants across multiple ecosystem partners. A larger network means more dependencies, more transaction volume and a wider potential surface for operational and fraud risks. The bank therefore needs systems capable of isolating failures, identifying unusual behaviour and degrading gracefully when individual components encounter problems.
Beyond lending
ANEXT Bank’s technology agenda also extends into mobile banking and onboarding, with eKYC and biometric verification forming part of its 2026 customer journey.
For Oh, deploying new technology requires looking beyond whether it performs under normal conditions. The more important questions concern who can use it, what happens when it fails and how the bank protects customers against security and fraud risks. “The question I ask isn’t ‘does the technology work’, but rather it’s ‘does it work reliably for everyone we serve, and does it fail safely when it doesn’t’.”
ANEXT Bank has reported more than 22,000 businesses now use its services. The Bank also showed 281% year-on-year growth in total loans disbursed and 207% growth in deposits as of December 2025.
For Oh, however, technology ultimately comes back to the people building it. Scaling the infrastructure requires an engineering culture capable of maintaining reliability as complexity grows. “What really keeps me up is how do I continue to build and motivate my team to deliver the best services to our customers.”
For ANEXT Bank, the next phase of SME lending will depend not simply on how quickly a business can receive a credit decision, but on whether banks can combine richer data with sound risk management and resilient technology. The shift from paperwork to live transaction data could make lending more responsive, but sustaining that advantage will depend on the systems and people behind it.
