Philippine BNPL improves risk segmentation and fraud detection with RiskSeal

Client's bio

This client is a fast-growing BNPL provider in the Philippines (company name confidential under NDA).

The company enables consumers to qualify for BNPL in minutes and use it instantly for everyday purchases. Customers can spend their credit with major merchants across the Philippines.

The provider focuses on financial inclusion. It helps underserved and thin-file consumers access essential goods such as food and medicine. Customers can build their credit history over time through responsible repayments.

Challenge

The company was growing fast. Approval speed was a competitive advantage, and customers expected instant decisions at checkout.

But behind the scenes, the risk team was facing a difficult reality.

More than 70% of applicants had thin or no credit history. Bureau data provided little insight, leaving large gray areas in the decisioning process.

Approving too cautiously meant losing good customers to competitors. Approving too aggressively meant rising defaults.

At the same time, fraud was becoming more sophisticated.

Synthetic identities and repeat defaulters were finding ways to re-enter the system.

With limited visibility beyond traditional credit checks, detecting these patterns early was increasingly difficult.

The team needed a way to move fast, without flying blind.

They needed deeper signals to separate real, reliable customers from risky or fraudulent applicants, especially in the mid-risk segment where most uncertainty lived.

Solution

The BNPL provider integrated RiskSeal into its credit decisioning flow.

Through a single API, the company added a new layer of alternative digital signals on top of existing bureau checks.

These signals incorporated unique data from high-usage local services and apps in the Philippines.

This included signals such as:

Email age

Phone number stability

Social media presence

Subscriptions

Web registrations

Results

RiskSeal helped the provider detect synthetic identities and repeat defaulters early in the decisioning process.

Key improved metrics include:

The biggest impact was in the mid-risk segment.

This group often sat between clear approvals and clear declines. Bureau data alone lacked the depth needed to confidently distinguish reliable borrowers from risky ones.

After layering digital identity signals on top of traditional credit checks, the company achieved a a 5% drop in default rates within this segment.

Many applicants who had been misclassified turned out to be creditworthy. It resulted in a 12% increase in approval rates.

As shown in the chart below, the enhanced scoring created a clear separation across risk bands.

Default rates decrease consistently as score bands improve - from high-risk segments to low-risk segments - demonstrating stronger predictive power.

This clearer segmentation allowed the risk team to confidently approve more applicants in higher score bands while tightening controls on lower ones.

The result was sustainable growth with improved portfolio stability.