Banking : Credit Behaviour Intelligence

Loan providers’ interest in credit behavior intelligence has significantly increased. In a way, this is not surprising given the rapid and unpredictable changes in business environments. Despite tremendous efforts to assess the potential risks of loan applications, it is impossible to fully account for unforeseen circumstances. For instance, the outbreak of COVID-19, wars, climate change, or political shifts—none of these disruptive issues were foreseeable. This is not a matter of AI methodologies, nor is it a problem of data quality or human insight. There will always be circumstances that no one could have imagined.

In reality, there are two major methodologies used to handle unforeseen circumstances. The first is developing credit behavior scoring models using loan transaction data. This approach helps to identify the likelihood of defaults or delinquencies after loan approval. Risk managers can proactively respond with strategies like demarketing to prevent opportunity loss. The second is controlling micro-patterns with business rules. Since the modeling process can take several months and credit scoring models require approval from financial regulators upon updates, business operations cannot solely rely on the model. Risk managers monitor and analyze the data, and if new delinquency patterns are detected, they quickly adjust approval strategies using business rules.

The difficulty of this task mainly stems from fragmentation. Well-informed decision-making is based on a comprehensive understanding of the business situation. However, credit scoring systems, rule systems, and scoring models and analytics tools like Excel, business intelligence and DS/ML platforms are not unified. Consequently, risk managers must repeatedly collect the necessary information from various sources. This is a highly time-consuming task. Furthermore, data analytics and machine learning require statistical or programming skills, so reliance on other departments is inevitable. 

DEIN Station's Credit Behavior Intelligence integrates these fragmented systems into one unified platform. By building a higher meta-intelligence that incorporates these intelligent sources, the overall business situation can be clearly assessed, and simulation, optimization, monitoring, and updating can be performed without limitations. Risk managers can observe the business status from a higher perspective, and the system can automatically alert them to unforeseen events or KPI fluctuations. For newly detected patterns, risk managers, sales and marketer, and product managers  can simulate the impact of their strategies and immediately update the system without separate development procedures.

Business KPIs 

• Total Loan Amount

• Loan Portfolio Yield

• Unused Credit Line

• NPL(Non-Performing Loan) Ratio

• Default Rate

• Default Amount

• Additional Loan Needs Amount

Challenges 

• Enhancing the profitability of the loan business

• Assessing default or delinquency risks using behavioural data 

• Simulating refinancing & cross-selling strategies 

• Identifying segments with potential additional loan needs 

• Monitoring and adjusting the strategies in a timely manner

Decision subjects 

• Refinancing

  - Interest Rate

  - Monthly Repayment Amount

  - Loan Term (contingent on interest rate differential and fees)

  - Rate Type (Variable or Fixed)

• Cross Selling Targeting