Industry: Fintech (Lending)

Location: Vietnam

Spokesperson: Zhan Zatayev, Chief Risk Officer, SFinance

Formerly known as SHB Finance, SFinance partnered with HyperVerge to transform digital lending through intelligent onboarding, automated decisioning, and built-in fraud prevention.

As digital lending continues to scale in Vietnam, financial institutions face a difficult balancing act: process more applications, make faster decisions, and strengthen risk controls without adding friction for legitimate customers.

SFinance, formerly known as SHB Finance, faced this challenge as its lending volumes grew. High application volumes, customer drop-offs, manual reviews, duplicate applications, and evolving regulatory requirements created pressure on both the customer experience and risk operations.

Together with HyperVerge, SFinance rebuilt key parts of its digital lending journey around intelligent automation combining identity verification, biometric authentication, NFC verification, deduplication, financial data analysis, and fraud prevention.

The transformation enabled SFinance to process 450K+ applications quarterly, achieve an ~80% approval rate, reduce manual review to ~1.7%, and maintain an error rate of less than 1%.

The Challenge: Scaling Lending Without Scaling Friction

SFinance is a consumer finance company focused on lending to Vietnamese individuals with government-issued national IDs.

As digital lending volumes increased, SFinance encountered several challenges that could affect both operational efficiency and customer experience.

• High application volumes
• 17–18% customer drop-offs
• A constant trade-off between manual review and approval
• The need to keep pace with evolving regulations

For the risk team, the challenge was particularly acute. Every suspicious application could not simply be sent for manual review. While manual checks could help manage fraud risk, excessive intervention could slow down legitimate customers and cause them to abandon the journey.

The objective was therefore not simply to automate lending. SFinance needed to make fraud prevention and risk assessment increasingly invisible to legitimate customers while maintaining strong controls behind the scenes.

The Transformation: From Manual Processing to Intelligent Automation

SFinance partnered with HyperVerge to rebuild its digital lending approach around three key areas:

01. Smart onboarding
02. Real-time underwriting
03. Built-in fraud prevention

1. Smart Onboarding: Building Trust from the First Interaction

The first step was establishing that the applicant was who they claimed to be and that they were a genuine, live person.

HyperVerge enabled SFinance to combine multiple identity and biometric capabilities within the onboarding journey:

OCR + Liveness + Face Match

OCR extracts identity information from the customer’s document, while liveness and face matching help establish the authenticity of the applicant.

Together, these capabilities create a streamlined first layer of identity verification.

NFC-Based Verification

With Vietnam’s national ID supporting NFC, SFinance began leveraging data stored on the card’s NFC chip as an additional source of customer information.

This data could subsequently support underwriting and decision-making, creating opportunities to move beyond identity verification toward richer customer assessment.

1:Many Face Deduplication

A major challenge was identifying whether an applicant was genuinely new or an existing customer.

Changes to national IDs meant that relying solely on the national ID number was not always sufficient to identify returning customers.

SFinance therefore introduced 1:Many face deduplication, allowing a customer’s face to be compared against its existing customer base.

This helped SFinance distinguish between new-to-bank customers and existing customers eligible for cross-sell opportunities.

2. Real-Time Underwriting: Moving Toward Smarter Decisions

Identity verification is only the first step in responsible lending.

SFinance was also looking at how it could better understand a customer’s financial position and use that information to support faster and more structured lending decisions.

As part of this journey, SFinance began piloting AI-based bank statement analysis with HyperVerge.

The solution is designed to help analyse customer income and expenses, creating a more structured view of financial behaviour that can support underwriting decisions.

This opens the door to more contextual underwriting particularly for customers whose financial profiles don’t fit traditional salaried employment patterns.

As Zhan highlighted, different customer segments may require different underwriting logic, including freelancers and customers with non-traditional income patterns.

3. Built-In Fraud Prevention: Preparing for the Deepfake Era

As digital lending becomes increasingly automated, fraudsters are also adopting more sophisticated techniques.

Deepfakes and presentation attacks create a new challenge for digital identity verification: an applicant may appear authentic without actually being the genuine individual.

SFinance addressed this by incorporating biometric fraud prevention capabilities into its onboarding stack.

HyperVerge’s solution provides ISO Level 2 Presentation Attack Detection (PAD) compliance, helping strengthen protection against sophisticated presentation attacks and deepfake-based threats.

The objective is not to add friction to every customer interaction.

Instead, fraud prevention should increasingly become invisible to genuine customers while remaining effective against fraudulent activity.

The Impact: Automation at Scale

The transformation has allowed SFinance to process substantial lending volumes while significantly reducing its dependence on manual review.

450K+Applications processed quarterly

80% Approval rate

~1.7% Manual review

<1% Error rate

These results demonstrate the impact of moving from predominantly manual processing toward an intelligent automation model.

Instead of manually reviewing a large proportion of applications, automation handles the majority of the journey, allowing human intervention to focus on cases that require additional attention.

The Next Frontier: Making Trust Invisible

For SFinance, the evolution of digital lending doesn’t stop at automation.

The next challenge is to make the experience faster and more seamless for legitimate customers while continuously strengthening fraud and risk controls.

SFinance’s roadmap focuses on three key pillars.

1. Invisible Fraud Prevention

Fraud prevention should happen without creating unnecessary friction.

The goal is to protect the institution while ensuring that legitimate customers can move through the lending journey quickly.

2. AI-Led Underwriting

Customers have increasingly diverse financial profiles.

AI-led underwriting can help lenders develop more contextual decisioning approaches and better understand different customer segments.

For SFinance, this means moving toward underwriting that can account for different customer behaviours, income sources, and financial circumstances.

3. A Unified Onboarding Stack

The future of lending will bring together information from across the customer, merchant, and lending ecosystem into a unified onboarding and decisioning framework.

This represents the next evolution of digital lending: moving from individual verification and decisioning tools toward a connected trust infrastructure.

Building Trust at Scale

SFinance’s transformation represents a broader evolution in digital lending: moving from isolated verification steps toward an intelligent, connected lending infrastructure.

With HyperVerge, SFinance has brought together identity verification, biometrics, NFC data, deduplication, financial data analysis, and fraud prevention to create a more automated lending journey.

The objective is ultimately simple:

Make lending faster for good customers, smarter for lenders, and harder for fraudsters.

As SFinance continues to evolve its digital lending stack, the focus is shifting from simply verifying customers to understanding customers, making faster decisions, and embedding fraud prevention into every stage of the journey.