Beyond KYC: Lenders turn to 'alternative' data to assess risk, detect fraud

Risk assessment firms are using data available publicly or obtained through user consent but is not available through credit bureaus to detect mule accounts, syndicates, and risky borrowers

Platforms that make decisions on risk levels are using alternative data sources to help lenders detect syndicates attempting to access credit from regulated entities as well as to enable faster, more accurate underwriting to new-to-credit (NTC) users.

 

Companies are also training models on multiple forms of alternative data such as location data, third party app usage, SMS data, payments transaction behaviour, and metadata on location of a user, among others, to further enable underwriting to NTC customers.

 

The detection of syndicates stems from analysing various forms of alternative data — such as multiple government IDs linked to a single mobile number — in high-risk regions with a history of fraud.

 

“What we look for are aggregated insights. For example, whether a user is linked to five numbers, three of which are connected to four other PAN cards. The idea is to derive insights from such patterns without revealing any personal data,” said Venkat Srinivasan, chief analytics and risk officer, Bureau.

 

The idea is to understand the relationships between different types of alternative data and the insights they reveal.

 

For instance, if several companies are registered at the same 200-square-foot office, it could be a strong indicator of fraud, experts said.

 

Ashok Hariharan, chief executive officer of IDfy, explained that such analysis can enable lenders to identify if a mule or a fraudster is further linked to a larger network or syndicate, operating out of high risk zones in the country.

 

An April 2025 report by the Fintech Association for Consumer Empowerment (FACE) said that 83 per cent of lenders in the country used both traditional and alternative data to underwrite credit for their customers.

 

The report noted that for small-ticket loans below Rs 50,000, income verification carries more weight than traditional credit bureau scores.

 

Such verification is typically done through bank statement data obtained with the user’s consent.

 

Companies said user data such as location, VPN usage to mask digital trails, altered camera settings to bypass KYC, and multiple loan apps on a single device, among other indicators, can also help improve fraud detection.

 

“For instance, if a person has cheque bounces which were reported or FIRs (first information reports) were registered, these can be relevant from the perspective of credit. One will not find it in (credit) bureau, but these are very predictive elements if you are entering into a credit transaction,” Hariharan from IDfy said.

 

He added that only publicly available data or information obtained through valid user consent — such as data from the Ministry of Corporate Affairs (MCA) portal, court records, and criminal databases — can be used for analysis.

 

The primary motivation for any lender to tap into alternative data is the lack of credit bureau information and predictive analysis if the person being lent to has the capacity to repay.

 

“We have around 1 billion identities such as a telephone number, PAN card, driving license, among others and around 600 million personas,” Srinivasan from Bureau explained.

 

Companies said that they continued to invest in capabilities to develop proprietary algorithms given the scale of data they currently process.

 

“The search itself is complex and requires a lot of proprietary algorithms to be developed. That we have invested a lot of time and effort into,” Hariharan said.

 

The report from FACE further added that reducing fraud risks and non-performing assets (NPA) rates, along with increase in loan approval rates were the most expected strategic outcomes for using alternative data.