Blog Security & Fraud Protection
8x8 brand photo people g1785617988 2

Beyond Automated KYC: How Video Identity Checks Catch What Algorithms Miss

Singapore is Southeast Asia’s leading financial hub, home to more than 1,700 licensed fintech firms and over 100 MAS-licensed digital banking, payment, and capital market operators. When a customer applies for a high-value mortgage, opens a wealth management account, or requests a large credit facility, standard verification isn’t enough.

The stakes are too high. Fraudsters know this, which is why high-value accounts are prime targets for synthetic identities, deepfakes, and credential theft.

In Singapore alone, scam and fraud losses reached S$456.4 million in the first half of 2025, with identity theft topping the list of scam types. Digital bank fraud claims climbed from 2.1% of all industry complaints in 2023 to 8.7% by August 2025. Across the wider APAC region, identity fraud surged 121% year-over-year in 2024, with deepfake attacks increasing 194% and synthetic identities becoming harder to detect at the document level.

For financial institutions, one successful fraud attempt on a high-value account can wipe out the margins from hundreds of legitimate customers.

That’s where video-based identity checks come in. By adding a live human verification layer to high-risk onboarding, financial institutions can catch fraud that automated systems miss while keeping the process fast enough that legitimate customers don’t walk away.

 

Why High-Value Accounts Need Stronger Verification

Under MAS Notice 626, non-face-to-face (NFTF) customer onboarding is never presumed low-risk. Financial institutions must apply additional checks to manage impersonation risk when verification is performed without in-person contact. Risk-based due diligence frameworks already require enhanced verification for certain account types. The question is how to apply it without creating friction that drives legitimate customers away.

Scenarios where video identity checks add the most value:

  • High-value accounts: Wealth management, private banking, large credit facilities, and commercial lending where potential losses justify extra scrutiny. With AI-driven wealth management platforms now serving over one-third of Singapore’s retail investors, the volume of high-value digital onboarding is growing fast.
  • Remote onboarding without branch access: Digital-first institutions that can’t rely on in-person verification need equivalent assurance. Singapore’s 18 licensed digital banks operate entirely without branches, making NFTF verification their only option.
  • Customers flagged by automated systems: When algorithms detect anomalies but can’t make a definitive call, human judgment fills the gap.
  • Regulatory requirements: MAS expects financial institutions to conduct internal assessments of the effectiveness of any identity verification solution in mitigating impersonation and fraud risks, and not to solely rely on external quality assurance standards of technology service providers.

 

When Legitimate Customers Get Blocked

Overly cautious algorithms create a different kind of risk. Research shows that 68% of consumers have abandoned onboarding processes they found too difficult. In Singapore, 78% of consumers say they would switch banks over inadequate fraud protection — but that sensitivity cuts both ways. A clunky verification process can push high-value prospects to competitors just as quickly.

For high-value applicants, the stakes are even higher. A rejected wealth management prospect doesn’t reattempt. They move to a competitor with a smoother process, and their lifetime value goes with them.

The cause is usually mundane: a biometric mismatch from aging, an ID photo taken under poor lighting, or a recently reissued document that the system doesn’t recognize. The automated system flags the anomaly, but can’t resolve it.

Without a recovery pathway, the institution loses a legitimate, high-value customer to protect against a risk that didn’t actually exist.

 

When Real Fraud Slips Through Automated Checks

The opposite problem is equally dangerous. Financial institutions report that 1 in every 20 verification attempts is fraudulent, a 21% increase from the previous year. In Singapore, the proliferation of AI and deepfakes has made impersonation attempts far more convincing, with 65% of Singaporeans encountering scam attempts at least once a month.

Sophisticated attackers target high-value accounts specifically because the payoff justifies the effort.

Synthetic identities, deepfake videos, and stolen credentials are designed to pass document-level checks that work well for low-risk applications.

Automated systems catch volume-based fraud patterns effectively. But for a targeted attack on a single high-value account, the attacker only needs to fool the algorithm once.

A trained verification agent conducting a live video session can spot behavioral inconsistencies, challenge the applicant in real time, and make judgment calls that no algorithm can replicate.

Read More: Reinforcing Trust: Tackling Fraud with Stronger Authentication in Finance

 

How Video Identity Checks Work

A live video session exposes things that automated KYC simply cannot see. For low-risk applicants, that gap doesn’t matter much.

For a wealth management prospect or a mortgage applicant, it’s the difference between catching a sophisticated attack and writing it off as a loss six months later. Here’s what trained agents surface during a high-stakes verification session:

 

Real-time Liveness, Not Still-Frame Liveness

Deepfake video can pass a passive selfie check. It struggles when an agent asks the applicant to turn their head at an unscripted angle, hold up a hand on cue, or describe an object in the room. Unpredictability is the defense.

 

Document Forensics on the Physical Article

OCR reads text. An agent can ask the applicant to tilt the document, watch how the hologram refracts light, check microprinting under zoom, and verify the laminate edge. Forgeries that pass image-based checks often fail this kind of physical scrutiny.

 

Behavioral Inconsistency Under Pressure

Synthetic identities are usually operated by people who don’t know the person they’re impersonating. A coached impersonator hesitates on basic questions: Which branch did you open your last account at? Who referred you? Algorithms can’t ask these.

 

Coercion and Duress Signals

High-value account fraud increasingly involves a real person being coached or coerced off-screen. Agents can spot eyes flicking off-camera, scripted phrasing, or another voice in the background. None of this shows up in document upload flows.

 

How 8×8 Video Interaction Powers Identity Checks

8x8 Video Interaction live video identity check session showing agent and customer views across desktop and mobile, with photo annotation, in-call chat, call reference details, and individual agent performance stats.
During a live video identity check, agents can annotate the customer’s camera feed, capture photos, and chat in real time – all from the browser with no downloads required.

The point of video for high-value onboarding isn’t to replace automated KYC. It’s to layer enhanced due diligence on top of it for the applicants who warrant it. 8×8 Video Interaction is built to slot into that tiered workflow rather than run as a separate process.

  • Triggered escalation from your onboarding stack: When risk scoring or account thresholds trip an EDD rule, the API or iFrame integration hands the applicant directly to a verification agent without breaking the session.
  • A documented verification trail: Each session can be recorded with full video and audio, and agents can capture photos, annotate documents, and tag the customer’s GPS location during the call. Combined with detailed call logs – including duration, shared images, and call ratings – your compliance team has a documented trail to support EDD file requirements. Under the PDPA, video recordings and photos captured during verification sessions constitute personal data. Organisations must ensure they have a lawful basis for collection, obtain appropriate consent, and limit retention to what is necessary for the verification purpose.
  • Agent management and visibility: Automatic call distribution routes incoming verification sessions to available agents, while the admin dashboard on 8×8 Connect gives managers a consolidated view of agent performance, call history, and customer data – so you can staff and manage your EDD queue with full visibility.
  • Administrative oversight: The admin dashboard provides aggregated call insights across all agents, including call duration and ratings, giving supervisors the oversight they need to maintain verification quality. Managers can add or remove agents and review individual call logs and shared images after each session.
  • Enterprise-grade security and global infrastructure: 8×8’s platform holds SOC 2 Type II, ISO 27001, HIPAA, and PCI/DSS certifications, with infrastructure across data centers in Asia, Europe, and the US to support data residency requirements. 8×8 also holds the CSA Cyber Trust Certificate, awarded by the Cyber Security Agency of Singapore for good cybersecurity practices. With local offices in Singapore and data centres in the region, session data can be stored in compliance with local data residency expectations. Session recordings and call logs give your compliance team the raw evidence to meet local regulatory standards.

The platform handles institutional volume without compromising the assurance level that makes EDD worth doing in the first place.

 

8x8 Connect admin dashboard showing video identity checks overview with number of calls made, average call duration, and average call rating, with a connected calls trend chart.
The 8×8 Connect admin dashboard gives managers a consolidated view of video identity check sessions, including call volume, duration, and agent performance ratings.

Read More: Customer Support in Financial Services with 8×8 Video Interaction

 

Strengthen Your Identity Verification with Video

High-value accounts deserve high-assurance verification. Video-based identity checks give financial institutions the tools to stop sophisticated fraud while keeping legitimate customers moving through onboarding.

With 8×8 Video Interaction, you get secure, browser-based video verification that integrates into existing workflows, creates compliance-ready audit trails, and scales with your business. So don’t wait, talk to 8×8 today about adding video identity checks to your onboarding workflow.

 

FAQ – Video-Based Identity Checks in Singapore

  • What are video-based identity checks?
    Video-based identity checks use live video calls between applicants and trained verification agents to confirm identity. Agents examine documents, conduct liveness checks, and assess behavioral signals in real time.
  • When should financial institutions use video verification?
    Video verification is appropriate for high-value accounts, remote onboarding without branch access, cases flagged by automated systems, and scenarios where regulations require enhanced due diligence. In Singapore, MAS Notice 626 treats all non-face-to-face onboarding as higher risk, making additional verification measures essential for compliance.
  • Is video KYC compliant with Singapore financial regulations?
    MAS does not prescribe a specific verification technology, but requires financial institutions to apply additional checks for non-face-to-face onboarding and to assess the effectiveness of any solution in mitigating impersonation and fraud risks. Video KYC, when implemented with proper controls, supports compliance with MAS Notice 626 and AML/CFT requirements. Organisations must also comply with the PDPA when collecting, using, and storing biometric and personal data captured during video sessions.
  • How does video verification detect deepfakes?
    Trained agents use interactive liveness checks, behavioral observation, and visual cues to spot deepfake attempts. The combination of human judgment and AI-assisted tools provides stronger fraud protection than automated systems alone.
  • Does video verification slow down customer onboarding?
    Most video verification sessions are completed in minutes. While longer than instant automated approval, video verification is much faster than email exchanges, document re-uploads, or branch visits.

Explore Related

8x8 Brand Photo Retail 1366720422
Technology & Transformation
April 29, 2026

WhatsApp Flows: Simplify Service Selection and In-Chat Ordering

Jason Chan

See how WhatsApp Flows replaces clunky forms with in-chat service selection and ordering. Compare basic vs. API-connected Flows and learn how to deploy at scale with 8x8.

8x8 Brand Photo Office G2196936957
Security & Fraud Protection
April 29, 2026

Verification APIs: A Complete Guide

Benjamin Kuo

A practical guide to choosing and integrating Verification APIs - covering OTP, silent auth, number lookup and APAC compliance.

8x8 Brand Photo Retail 1472087329
Security & Fraud Protection
April 28, 2026

Resolving Failed eKYC Checks with Live Video Interaction

Igor Mostovoy

When automated eKYC flags a real customer, you need a recovery path - not a dead end. See how live video verification turns failed checks into completed onboarding.

Talk to an Expert

Complete this form and an 8x8 sales specialist will reach out to you shortly.

8x8 SMS API allows you to efficiently reach your customers by automating notifications, OTPs, reminders and alerts with a high delivery rate.

Thank you for your interest!

An 8x8 sales specialist will reach out to you shortly.

Any urgent enquiries and help needed?

Email [email protected]

Support Visit support site

To top
This site is registered on wpml.org as a development site. Switch to a production site key to remove this banner.