Automating decision making

A FinTech client specializing in online business loans needed to scale a decision engine that made quick decisions about credit risk while allowing human credit analysts to easily customize the credit scoring process. As the clients’ partner banks began to integrate into the platform, increasing the complexity of business logic became a challenge.

Our solution: build a rules management system IDE for credit analysis with a user interface to manage rule sets and visualize the decision-making tree. Without help from tech experts.

A few of the system features:

• Visual editor that shows a tree representation of the decision-making process and allows the user to create new steps. Among other things, it allows the system to create tasks for the credit analysts and process their input.
• Exportable and testable policies, the root of any decision-making process. The average exported policy file consists of about 4,500 lines of code, or a few hundred decision-making logic steps.
• Visualization that allows analysts to quickly see why a decision was made, at what step the application was declined, or what documents are missing, why the loan amount is so small, and so on.

We built a robust solution for our FinTech client: a fully managed decision engine and a process visualization system that supported thousands of rules across partners and enables credit analysts of all technical abilities to easily understand and interact with the platform.