Key problems facing financial institutions:
- Complex Traditional Banking Structure: It is imperative for banks to improve their traditional cost structure in order to maintain its competitiveness in the market. Also, the adoption of new technologies is impeded, mainly due to their legacy IT system, organizational complexity, and scarce of AI talents with financial expertise.
- Operating Inefficiencies: Digital banks, on the other hand, offer more competitive products since they have a flexible structure with no legacy system. Yet, decision-making processes in core banking functions are monitored and managed separately, leading to inefficiencies from duplication of process and IT investments. It is difficult to achieve economies of scale in operations despite having a smaller and simple structure.
- Low Usability of Finance in Data-rich Platform: High barriers to entry to offer lending services using their e-commerce and telecom data.
Massive-scale Machine Learning Automation Platform
Our solution is specialized in Banking, Payment, Insurance and Asset Management. With its high scalability and ease of adoption, The solution is being applied to every decision-making framework from product development, credit scoring, underwriting to up/cross-selling, risk management and fraud detection, etc.
Our solution also serves as a core AI 'Logic Chip' that integrates and converts cross-industry data into 'credit' data in Finance. We are the first company to be authorized by Korea's Financial Services Commission(FSC) to conduct AI loan underwriting on behalf of a U$400B bank. Our company is now redesigning digital banking by offering modularized Banking-as-a-Service(BaaS) to non-banks like e-commerce, telecom and e-wallet providers in the Southeast Asian market.
Technology Features, Specifications and Advantages
Our solution automates the full end-to-end modelling process from Data Collection, Pre-processing, Feature Engineering, Model Training, Validation to Real-time Prediction. Even people without AI expertise can easily build, monitor, deploy and update thousands of predictive models with just a few clicks.
ABACUS Key Highlights:
- Massive-scale data processing and cross-industry data integration with in-memory processing
- Modelling automation with real-time model performance validation and updates
- Explainable AI(XAI) with human-interpretable models, even under Deep Learning algorithms
- Domain-specific AI business applications across the value chain
- Seamless integration with legacy systems and rapid model deployment using REST API
Our solution is widely applicable to every decision-making process across the value chain in Banking, Payment, Insurance and Asset Management. Some examples of AI business applications include:
- Underwriting (personal/SME loan)
- Fraud detection in retail payment
- Multi-dimensional credit-decision on e-wallet
- Custom loan/micro-insurance product development
- Product recommendation using cross-industry data
- Target marketing (up/cross-sell)
- Automated credit cycle control module
- Asset sale and capital optimization
- Banking-as-a-service for e-commerce, telecom, etc.
It is now operating on U$10B personal loan assets in one of the largest banks in Korea and is now pioneering in deep learning applications in detecting fraudulent payment transactions in real-time. Using the same framework, our solution is expanding to health care and the public sector with high scalability.
1. Financial Institutions
Financial institutions can radically improve the traditional cost structure (both credit and operating costs) from economies of scale using our solution. They can maximize operational efﬁciencies from massive-scale modelling automation, with potentially new revenue opportunities created from differentiated strategies.
From the user perspective, business domain experts who want to leverage data and domain knowledge to make data-driven decisions can easily make use of the solution, even if they lack AI expertise and coding skills. Our solution also simplifies the most time-consuming tasks for data scientists or analysts by automating the end-to-end modelling process with simple user interfaces. They can easily build and monitor thousands of models, and deploy the most optimal model in real-time with just a few clicks.
2. Individual Customers
Banks traditionally analyzed customers based on a single measure (e.g. credit risk), which has limited the diversification of product and service offerings for individual customers. With the rise of open banking and cross-data integration, customers can benefit from our multi-dimensional models via customized products (e.g. lower interest rates, higher credit limit, etc.) and more affordable financial services.