Banking on bots for loan recovery

Keeping Tabs On Delinquencies

- NPAs make a big dent in the loan books

- NPAs in the 2-wheeler segment grew by 140bps

- 30bps in loan against property segment

- 11bps in the personal loan segment

- Mostly borrowers forget EMI date or cash withdrawal to pay EMI

Mumbai: Debt recovery bots have redefined loan collection processes for banks and NBFCs. After nearly two years of ups and downs due to Covid-19, the Indian economy finally is showing signs of revival. An important indicator of this revival is the strong demand for retail loans.

However, banks and NBFCs have another problem to tackle, i.e., rising delinquencies on two-wheeler loans, loans against properties, and personal loans. As per the data released earlier this month by credit monitoring agency Transunion CIBIL, NPAs in the two-wheeler segment grew by 140 basis points, in loan against property segment by 30 basis points, and 11 basis points in the personal loan segment. These NPAs make a big dent in the loan books of lenders and cripple the banking ecosystem.

While many loans fail because of significant reasons like loss of income, fraud, and lack of intent to pay, it's surprising that many of them fail due to mundane reasons like borrowers forgetting the EMI date or cash withdrawal to pay the NBFC, etc. Traditionally, the lenders would appoint a loan recovery agency that used muscle power or other ugly means to recover the money. The entire process would severely affect the relationship between the lender and borrower forever. After the RBI's strict regulations, this evil practice has stopped considerably, leaving borrowers to find other ways to manage delinquencies.

The good news is that the financial institutions can automate the pre and post-delinquency management cycles with AI-powered contact centres such as Rezo.ai. Financial institutions send reminder text messages and emails to borrowers at the pre-delinquency stage. This approach doesn't cut across sections in a country where literacy rates are low and people speak hundreds of languages and dialects. Rezo.ai AI-powered contact centre can make automated calls to borrowers to remind them about their EMI dates.

This call can be customised in different languages and dialects to ensure that the borrower understands the message and does not have mundane reasons for non-payment. This is not a one-way call, but the conversational AI capability allows the borrower to respond like a regular phone call.

The automated speech recognition powers the system to assimilate the audio data, convert the speech to text with high accuracy, and run sentiment analysis, as well as empathy and tone analysis. This analysis can be constructive in predicting the intent to pay. If there are red flags, the financial institutions can take additional measures to ensure a timely repayment.

Talking to Bizz Buzz, Dr Rashi Gupta, chief data scientist & co-founder, Rezo.ai, says: "It is vital for India's socio-economic progress to disburse retail loans steadily. At the same time, it is essential to ensure the safety of the overall banking ecosystem."

To a great extent, AI-based contact centres can resolve this problem. These systems can be customised to meet a financial institution's specific delinquency management requirements. The best way to start a digital delinquency management journey is to identify a technology partner that can deploy an AI-based contact centre for financial institutions, she added.

The AI-powered contact centre can also be accommodating through the post-delinquency management stage. At the back end, the system generates insights based on deep-data analytics. These insights are valuable in creating meaningful nudges for borrowers. For example, the system can suggest which borrowers resume after missing 1-2 EMIs and those who are likely to default. As a result, those who start repaying do not need a hard nudge (home visit). Instead, they can be given automated calls to keep track of their sentiment. For others, lenders can plan additional measures. This ensures cost optimization and better recovery rates.