Client Payment Classification based on Natural Language Processing Qorus-Infosys Finacle Banking Innovation Awards 2025

The project’s necessity arose from the necessity of better understanding our client’s usage of the “BB Pay” checkout service. BB Pay is formally a digital solution for easy and personalized payment checkout, as a company can create a centralized payment link for a product or service and then share It with a client instantly. Therefore, a client can safely perform a payment given previous authentication in the Banco do Brasil (Banco do Brasil bank) mobile application or other banking apps. As a storeowner or service provider, BB Pay offers a straightforward checkout solution that lets the client easily track all payments being received of a given payment link without having to use a credit card terminal. The result is more security and transparency for the client that is making a payment. BB Pay currently provides its services for thousands of active monthly users in 2024 (among individuals and legal entities) and serves as an internal checkout for other services provided by Banco do Brasil bank. This solution also fits in different business models such as a financing and a physical contactless checkout option for stores. Currently, BB Pay's main clients are not only internal clients (such as checkout for BB's own services) but also the government, so there is a need to understand emerging clients. Given the data, this project aims to classify clients given their behavior using a mix of text-based data alongside traditional data, considering both LLM-based and traditional Machine Learning Approaches.