.png){kind=logo}
The evolution of lending in India tells a fascinating story of how fraud vectors have transformed. From the controlled environment of bank branches where customers physically presented themselves to today's embedded finance ecosystem, where loans are disbursed through super apps and digital marketplaces, the attack surface for fraudsters has expanded dramatically.
Banking began with branches, where face-to-face interactions formed the foundation of trust. Then came the era of Direct Selling Agents (DSAs), taking banks to customers' doorsteps. While this expanded reach, it also introduced new risks through human tendencies for collusion.
The web revolution promised greater access but opened Pandora's box of digital vulnerabilities. Today, with embedded credit, open protocols and networks like OCEN (Open Credit Enablement Network and ONDC (Open Network for Digital Commerce), lending has become truly ubiquitous - available at every digital touchpoint where consumers transact. But each touchpoint opened the floodgates of fraud.
While regulators have mandated robust KYC processes, the current framework has significant gaps. The probability of the reviewer approving an e-KYC application depends on how convincing the customer’s digital identity appears on-screen.
Old and poor-quality images used in identity documents like Aadhaar cards, PAN, or driving licences make a reviewer's job even more difficult and often leave adequate space for errors, meaning a convincingly fake document may pass the test. It is an open secret that Aadhaar card photos seldom match the face of that person. So, it heavily depends on the person conducting the e-KYC. Video KYC, meant to bridge this gap, can be fooled with deepfakes; the higher the quality of deepfakes used in verification by attackers, the higher the chances of success. Official documents like salary slips and bank statements can be easily fabricated using readily available tools.
And the cost is staggering. Cyberfraud losses could amount to 0.7% of GDP, according to a projection made by the Indian Cyber Crime Coordination Centre (I4C), which runs under the Union Ministry of Home Affairs (MHA).
These losses don't just impact banks — they translate into higher interest rates and stricter lending criteria, ultimately affecting genuine borrowers.
No single test is foolproof. Imagine a detective solving a murder case. They never rely on a single piece of evidence — not just the fingerprints, not just the CCTV footage, not just the witness testimony. It's when multiple independent pieces of evidence align — the DNA matches, the phone records place the suspect at the scene, and the financial transactions show motive — that they can build a solid case. Each piece strengthens the others. It’s a classic case of triangulation.
In lending, we can apply this principle by cross-referencing multiple data points to establish the authenticity of a claim. Let’s take the case of personal loans, a segment that has had rapid growth. Here are a few facts: 14% of borrowers falsify employment status on loan applications, 12.2% of Early Payment Default rate are linked to suspicious employment pattern, 18% of applicants use disposable emails effectively worthless.
Traditional verification involves calling HR departments or requiring multiple documents, creating friction in the customer journey.
Smart triangulation offers a better way. By using just data inputs of an applicant a phone number and company email you can cross-reference multiple independent signals:
Is the employment record active in EPFO?
Does the company email domain match the claimed employer?
Does the phone number's age and usage pattern suggest legitimacy?
Is the company email domain ownership sketchy?
Is the company email available and accessible?
Each signal alone can be faked, but faking all signals simultaneously, making them perfectly consistent, is exponentially harder.
Triangulation can apply across loan products. For business loans, triangulating GST returns with bank statements and supplier invoices can flag revenue inflation. For vehicle loans, cross-referencing chassis numbers with insurance databases and RTO records can prevent multiple financing.
As lending continues to democratise, the importance of robust yet invisible fraud prevention mechanisms grows. Smart triangulation represents a shift from document-based verification to intelligence-based validation. It's about connecting dots across disparate data sources to build a web of trust, making it exponentially harder for fraudsters to game the system.
The key lies in getting two things right: One, choosing the right data points - those that are hard to manipulate, independently maintained, readily accessible, and privacy-compliant. Two, utilizing platforms that seamlessly connect to a range of data sources for real-time verification.
Authored by Rajat Deshpande, Co-Founder and CEO, FinBox
[Disclaimer: The views expressed are solely of the author and Analytics Insight does not necessarily subscribe to it. Analytics Insight shall not be responsible for any damage caused to any person/organization directly or indirectly.]
