Infosys delivered a holistic credit risk analytics and management engine by combining IP assets and technology products of leading fintech enterprises. We established an Infosys Collaborative Connected Credit COE to implement best practices and distil actionable insights for improving risk management.
We created a machine learning model to understand the reasons for payment default and assess the credit worthiness of clients. The model provided better visibility into credit exposure across the risk spectrum. In addition, it identified potential risks across the receivables portfolio as well as potential clients to mitigate risk exposure via debt factoring.
Our credit risk assessment model applied a collaborative membrane to unify data in functional silos such as business development, sales, marketing, and manufacturing. It enhanced the accuracy of default analysis, thereby improving forecasting and receivables management.
Early warning system
Our machine learning model correlated payment terms, invoice exposure and credit risk factors across the supply chain by analyzing –
- Unpaid invoices over diverse timelines
- Days Sales Outstanding (DSO) for each payment term
- Credit limit assigned vis-à-vis outstanding balance for each client
- Credit limit utilization across business segments
- Net working capital cycle across debtors, inventory, creditors, and cash
- Pre- and post-shipment terms for export finance
- Realization of export trade flows
The analysis revealed patterns in customer ordering and payment behavior as well as credit exposure for clients, products and delivery. In addition, it identified the countries, products, payment terms, and clients with the highest default risk.
The model discovered significant mismatch between some payment terms and product categories, which caused a spike in default. It also revealed that credit limit was not specified for several existing clients and a significant number of new clients, which resulted in excess credit and higher default risk.
Credit risk forecasting
The Infosys credit risk analytics engine analyzed internal and external data of the past four years, including Letters of Credit, bank guarantees, commodity rates, exchange rates, and sales orders. Our risk management algorithms leveraged historical data to generate forecasts for credit performance and accounts receivables.
Our model helped determine factors contributing to default behavior and predicted default risk across clients, products and regions. It simulated future transactions and created client credit profiles, which helped standardize payment terms.