Dynamic Loan Evaluation Challenge
In my previous post, I addressed the SaaS Pricing Challenge from the DM Community. In this post, I’ll address the Dynamic Loan Evaluation Challenge which can be found here. This scenario is applicable to many of our customers. While the business logic is very simple, data is uncovered over time. As a result, the underwriting decision changes each time new information comes to light (hence dynamic). Overall, this project illustrates the combination of a point decision, the loan origination, and a changing series of facts. Click here to watch the demo
Key Takeaways from This Challenge
In this challenge, eligibility is based on one measure of risk exposure. If the assets exceed the obligations, we approve the loan. Conversely, if obligations exceed assets, we decline the loan. As we uncover more assets and more debts, the arrow switches between approve and decline, making the interaction seemingly dynamic. However, in real systems, eligibility criteria is more complex. Therefore, I prefer to design loan origination systems as stateless services. This means that you expose all available information and then apply the business rules to render the verdict. In short, the decision service does not keep track of the history (nor does it need to). In this regard, the evaluation itself isn’t dynamic.
However, the data collection is dynamic and the loan origination system needs to be able to ingest new data. I handle this with our Dynamic Questionnaires. The questionnaire collects all bower and guarantor information over time. As the loan engine, I can append the new facts to the in-flight application and then submit the appended application for loan evaluation. In other words, I prefer to cleanly separate the evaluation logic from the data collection logic.
If you want to try building the demo yourself with SMARTS™, feel free to ask for a free evaluation.
Learn more about Decision Management and Sparkling Logic’s SMARTS™ Data-Powered Decision Manager