How technology can strengthen fraud risk management in financial institutions

Take a gander at this data—The latest annual report by the Reserve Bank of India (RBI) says that as many as 36,075 frauds were reported in banks in the financial year 2023-24 (FY24), growing nearly 166% from the 13,564 cases reported in FY23.

The amount involved in fraud was as much as ₹13,930 crore. This is just from banks; if we include those from other financial institutions, such as non-banking financial companies (NBFCs) and housing finance companies (HFCs), the numbers will likely balloon. These numbers point towards significant losses not just to customers but also to financial institutions.

There are broadly two kinds of frauds that are happening today–one that defrauds the general public and another in which financial institutions themselves are defrauded–both involve monetary losses for the respective parties.

The first type of fraud is mostly perpetrated by unscrupulous fraudsters and involves social engineering. In such instances, it is usually the carelessness of individuals that results in them getting defrauded. The only cure for this problem is more awareness and keeping all necessary details related to money safe and secure.

The second type of fraud that involves financial institutions happens largely due to oversight and negligence in vetting the credit profile or documents provided by the borrower. It may also involve corrupt officials from the banks.

An evaluation of bank group-specific fraud cases spanning the past three years by RBI reveals that although the private sector reported more frauds, public sector banks consistently accounted for the highest total fraud amount. The majority of frauds, in terms of number, have occurred in the digital payments category (card/internet), while in terms of value, they have primarily been reported in the loan portfolio.

In any secured loan, borrowers have to provide collateral. Usually, the process is like this: a borrower submits the application to avail the loan. The bank then looks at the credit profile of the borrower, their history, and financial condition. Banks will ask for details of collateral for the loan. Then the bank will vet the documents before approving the loan.

Most frauds happen at this stage—which is called onboarding—as borrowers can lie or submit fraudulent documents. Due to workload and steep targets, bank officials have little time to vet every document carefully. Frauds can also happen once the loan is approved. For instance, the borrower can sell off the collateral without informing either the bank or the buyer.

Both scenarios are easily solvable, thanks to technology. Banks should employ advanced technologies such as artificial intelligence (AI) and machine learning (ML) at the time of onboarding, as well as for real-time monitoring of the collateral throughout the loan cycle.

Using technology while onboarding to vet documents can not only remove human errors but also make the task much more efficient. Moreover, such technologies can also bridge the language barrier —as each state’s property records are written in their respective local languages, making the task even more difficult and time-consuming for bank employees not from that particular state. Technology can also reduce challenges in monitoring of collaterals, which is humanly impossible to do for the entire portfolio of any large bank.

One example is monitoring collaterals submitted in farm loans against agricultural land. In some states, farmers are not even mandated to deposit land ownership papers with the bank if the amount of the loan is below a certain threshold. In such cases, it becomes extremely difficult to track any changes in land ownership after the loan has been furnished. AI-leveraged technology can pull publicly available property records from the state’s revenue department at set periods and then compare that against the data from the previous period for any changes. If there is any change in, say, land size, ownership names, etc., this can be flagged to bank officials who can raise queries with borrowers.

The RBI circular issued in June 2011 also mandates banks to implement a system where concurrent auditors examine and report on the authenticity of title documents, especially for large-value loans. Recently, RBI issued three revised Master Directions on Fraud Risk Management for Regulated Entities —one each for Commercial Banks, Cooperative Banks, and NBFC & HFCs—making directions clearer.

The real-time monitoring of collaterals will not just reduce instances of fraud but also alert the bank to any possible slippages ahead of time. This will also provide banks with the time to take legal action, avoiding the flight risk of borrowers.

Using such technologies also reduces human interference, mitigating the risk of corrupt officials causing losses to financial institutions.

The RBI has been proactive in announcing directions from time to time to mitigate fraud in the banking system. The recent master direction that collates all circulars and directions into one framework makes it easier for financial institutions to follow. This also reduces the number of compliance requirements for them.

Apart from making their credit system robust by complying with every direction from the RBI, financial institutions should also run awareness campaigns so that borrowers are sensitised and do not fall into any traps. They should also be made aware of the consequences of furnishing incorrect documents.

(Vishal Sharma is the Co-founder & CEO of AdvaRisk, a fintech startup)