Document Verification in the Philippines: AI Limitations and Opportunities

Online platforms are part and parcel of our lives in the Philippines — from online banking to ecommerce. A crucial time to engage with these services is during digital onboarding; the initial phase when users prove their real-world identities matches their digital personas.

This stage often sets the tone for the user experience. In fact, according to the ABBYY State of Intelligent Automation Report, 90% of companies lose potential customers during this phase.[1]

What is document verification?

Document verification typically occurs at the start of onboarding. Users present a government-issued ID like a passport or a UMID card — something many Filipinos are familiar with. Advanced optical character recognition (OCR) technology extracts relevant details from the uploaded document, saving users the hassle of manually typing them in.

Artificial intelligence (AI) further scrutinizes these documents for authenticity, checking for security features like holograms, watermarks and foils. In rare cases where the algorithm fails to pick up details, a fallback process is triggered to enable manually captured identity details.

[1]  90% of companies lose potential customers during the digital onboarding process, according to ABBYY State of Intelligent Automation Report Q4-2022 | Business Wire

Selfie check and liveness detection

After the document passes the AI check, the process usually requests a selfie for further verification. But a simple selfie isn't enough because a fraudster could use a photo of a legitimate ID holder.

That's where liveness checks come in. TransUnion's passive liveness check adheres to the ISO 30107-3 Level 2 standard, the industry's highest. This level confirms the user is blinking and not asleep, and screens against advanced fraud attempts, including realistic dolls, 3D masks and digital deepfakes.

AI limitations

Relying solely on AI for identity verification has its downsides, and it's essential to understand the limitations. AI models are only as accurate as their data, and they can misflag documents as either fake or genuine. AI also struggles to adapt quickly to new types of fraud, especially those not yet part of its training data. In the Philippines, where identity theft and vishing, smishing and phishing are on the rise, these limitations can be costly. 

Some companies augment AI with manual reviews, particularly in emerging markets. However, this can slow down the process, adding friction for the user and increasing drop-off rates.

The TransUnion approach

TransUnion Philippines goes beyond these traditional methods with a defense-in-depth approach to tackling identity fraud. Apart from OCR and facial recognition, our Digital Onboarding Suite verifies IDs against trusted local data sources like the Philippine Statistics Authority databases and local credit bureaus. This multilayered methodology provides a more reliable, faster, user-friendly digital onboarding experience.

The bottom line

TransUnion Digital Onboarding Suite offers a robust, holistic approach to secure digital identity verification. We go beyond AI's limitations by integrating document and facial recognition with advanced liveness checks and additional verification against reliable data sources. All these features work together to not only enhance security but also improve the user experience and help drive business growth.

The best part?

These capabilities are just one API call away. To unlock enriched security and seamless user experiences, email us at DigitalSuite@transunion.com today.