Unmasking fraudsters: real-time data changes the credit game

Source: Supplied. Michael Bowren, co-founder of Finch Technologies.

Fraudsters are becoming more sophisticated, rapidly adapting to outsmart the very technologies designed to combat them. In a sprint to stay ahead, real-time user risk profiling and the adoption of technologies designed for this will prove to be the answer to combating falsified customer data.

Evolving from historical limitations to real-time solutions

For many companies, their risk profiling still primarily relies on historical data to assess a customer's creditworthiness. User profiles are a data-driven narrative that provides insights into spending habits, income, expenses, and behaviour.

It can be as detailed as including a customer’s geographical location, age, and gender, helping businesses assess risk and tailor their services for a more personalised customer experience. However, traditional risk profiling has its limitations. It often relies on segmented, backwards-looking data - such as three-month bank statements which can miss red flags.

A perfect example of this - a lender may see a customer as financially stable based on consistent cash flow in their bank account. However, on closer inspection might reveal the customer is engaging in salary mimicry. Similarly, relying on credit bureau reports to assess open accounts and credit facilities can be misleading. A borrower may appear to have a manageable credit profile, yet they could have secured funding from three different lenders within 24 hours.

The rise of regulatory demands like KYC and AML compliance has amplified the need for advanced fraud=prevention tools. This is where real-time risk profiling truly changes the game.

By analysing live transactional data, businesses gain an immediate and accurate view of a customer’s financial behaviour. In analysing 100,000 South African consumer bank statements, Gathr, a data collection and verification platform, discovered that 14% of bank statements were fraudulent.

Pam Golding Properties data breach: What you need to know and what’s being done

4 hours

By examining metadata and cross-referencing data, platforms like this can allocate a fraud score to a customer, which determines the likelihood of a fraudulent event.

There are two major types of fraud driving the need for real-time profiling:
First-party fraud – This occurs when an individual manipulates or falsifies documents to appear more creditworthy than they are, or they can be impersonating another individual who has a better credit rating.  
Second-party fraud – This happens when a person uses their real personal information and legitimate documents but engages in deceptive practices, such as income mimicry, to secure credit or financial services. 

Impersonation fraud in South Africa has surged by 337%, with criminals using stolen identities to open or take over accounts, threatening financial institutions and their clients. Over the past two years, a significant portion of these attacks targeted National ID cards, which accounted for 80% of all document fraud incidents. This type of ID fraud can be a gateway to income falsification, as fraudsters use another person’s ID and credit rating to appear financially stable with a steady income.

Companies adopting new risk detection

In a direct effort to address these risk challenges, tech companies are making fraud prevention a core focus, integrating advanced risk detection strategies into their product roadmaps to stay ahead of emerging threats for their clients.

Insurers like Naked use AI to analyse customer behaviour during the application process, identifying patterns typical of honest customers. While most insurance companies rely on 20 to 30 data points for risk assessment and payout decisions; Naked cross-references and analyses thousands. This advanced approach allows them to fast-track approvals for genuine customers while efficiently flagging potential fraudsters.

South African tech providers are the champions behind this shift. Platforms like Gathr streamline the entire process by automating document verification, cross-referencing data, and using open banking APIs to provide real-time access to banking data.

Moving toward predictive intelligence

Being five steps ahead of fraudsters requires adopting real-time systems that cross-check diverse data points, such as active bank accounts, credit reports, and recurring payments. Fraudsters, although they are usually using technology, are still human, which means they inevitably make mistakes when falsifying data, and the key to preventing fraud lies in catching these slip-ups at every stage of the customer journey.

For industries like telecommunications, predictive intelligence has enormous potential. Companies such as MTN and Vodacom could use these technologies to foresee defaults on cellphone contracts, analysing real-time data for anomalies like irregular salary movements.

Real-time risk profiling is set to evolve into a predictive tool, not only preventing fraudulent events as they occur but also forecasting potential fraud by analysing patterns from similar customers.

Imagine having the power of a dashboard where a company could punch in a customer’s ID number and instantly check anything from their likelihood to repay a loan to their risk of committing fraud before it happens - all based on previous patterns. This predictive capability helps businesses protect their bottom line, while for consumers, it means greater security and better customer experience.