An Advanced BNPL Solution for an eCommerce Business

The project is a solution for “buy now pay later” offers. It belongs to the type of short-term financial products that allow consumers to make purchases in stores and pay for them later.

The application offers a credit option at the checkout of partnered online stores that make it possible to pay the total over a period of time. To use the app, the customer performs a simple registration procedure and selects what they would like to buy from the partner stores. After that, users reach the check-out stage to pay later, with installments and terms fitting them best.

To make it safe, a 2-step approval decision-making module was integrated. It functions on an automated and manual basis. The automated one makes purchases simple for customers, owing to the absence of challenging credit record checks. At the same time, the manual mode helps responsible managers approve and decline some specific purchase requests.

Reminders and notifications module helps customers stay updated on their activities and purchases made via the application.

There are no hidden fees when one pays in a timele fashion. The idea behind the fees module is to help responsible managers to control the size of the fees to be paid in the case, for example, late payments. As soon as the store ships the order, the user receives an email from the application outlining the payment schedule details.

The Buy Now, Pay Later (BNPL) application was implemented with the use of microservices and event-driven architecture.

The application was implemented via microservices and event-driven architecture.

Various application modules expose corresponding functionalities to external consumers via APIs, with each API created as a container-based microservice. As for the APIs themselves, they are exposed to external consumers via the API gateway through which the authentication and authorization procedure is controlled. Finally, business processes within the modules are managed and orchestrated by the Business Process Management Engine.

Each of the Modules uses NoSQL Database for transactions and operational purposes. RDBMS is applied to manage structured data. Data caching is also performed, where applicable.

The Next Best Offering component of the Decision Making Module is built on the basis of ML capabilities.

Modules are completely decoupled and interact with each other via the Event Streaming Platform.

The first person is responsible for creating the request (they are known as users), while the second person checks and approves such actions (they are known as administrators). This module helps specify a limit for the purchase order amount for the user. If the user tries to make a purchase order beyond the limit, this purchase order will automatically be directed to the 'To Approve' stage. That is where the admin can approve or refuse to approve this purchase order from 'Purchase Order Approval'. This module can help track approved or refused orders by such properties as "approved by", "refused by", "approval date", "refusal date", "refusal reason". It is possible to notify and alert about approvals via an email sent to a specific user or the manager responsible for approving a purchase limit.