4Paradigm’s ‘Sage’ Intelligent Risk Control Platform
An Intelligent Decision-making and Risk Management Platform to End-to-end Solve All Kinds of Fraudulent and Credit Risk Problems

Product Value

Highly Effective

It utilizes independently developed high-dimensional machine learning technology combined with traditional expert rules to help customers achieve an overall improvement in the accuracy and coverage of risk-controlled anti-fraud business.

Comprehensive Capacity

Relying on industrial-leading machine learning, decision engine, natural language processing, complex networks, knowledge graph and other technologies to build an intelligent AI-level anti-fraud brain.

Cold Start

Built-in general risk control rules and models summarized from 4Paradigm’s industry experience, helping financial institutions quickly and easily establish a standardized risk control process.

High Performance

High throughput, low latency service capability, millisecond response support, outstanding model accuracy, making real-time risk interception possible.

Independent and Controllable

Localized deployment of intelligent risk control platform, independent core risk control capabilities; cloud services support, faster and lower cost and more efficient expansion of AI risk control capabilities.

Easy to Use

The program has been launched at a number of large banks, with fast implementation and outstanding results, equipped with general rules and models to support cold start and able to incrementally upgrade and continuously evolve.

Product Functions

  • Unified Model Center Management
    User-defined Modeling

    Whole-process model research capabilities from data import, data cleaning, data segmentation, automatic feature extraction, model tuning, model evaluation, etc.

    Model Self-learning

    The model center can use the new data generated from the business scenario to perform model iteration in real time to prevent the model effect from decaying and keep up with business development changes.

    Model Online Prediction

    With the self-developed time series in-memory storage engine, feature calculation engine and estimation engine, the model can be launched into operation with one click, providing real-time model estimation service.

  • High Performance High Availability Decision Engine
    Decision Flow Management

    The decision flow is edited and managed through the visual operation interface. The rules and models can be flexibly accessed, and the operation is simple. It can be easily used to support the decision flow hot deployment.

    Rules Center

    Built-in multiple logical operators, function calculation formulas, and time series data processing methods, the rule configuration can be completed without manual coding, and the decision flow can be directly invoked.

    Universal Solution

    Built-in general risk control rules and models of 4Paradigm’s industry experience, helping financial institutions quickly and easily establish a standardized risk control process.

  • Case Management Close to Business Scenario
    Case Information Inquiry

    Based on artificial intelligence technology and deep mining, it can inquire and analyze all transaction/application information of the customer's history, and conduct detailed analysis, confirmation and summary of each case through standard case processing.

  • Indicator Center
    Indicator Center Close to Business Scenario

    The indicator center provides an indicator system including business indicators, model indicators, and monitoring operation and maintenance indicators to facilitate users to track and monitor business results and analyze business trends. Unique custom indicator calculations are closer to the customer's business. Also, the indicator center provides a visual representation of indicators, which improves the efficiency of viewing, communicating and presentation.

  • Multi-dimension Business Monitor Dashboard and Risk Report Display
    Business Monitor

    Overall monitoring of business from the dimensions of transaction volume, interception volume, actual fraud amount, and rules hit rate including the data comparison of today, yesterday, last week, and last month etc.

    Report Management

    Output report from a variety of dimensions, including but not limited to, rules, transactions/applications, cases, circulation, models, etc., to meet financial institutions' daily, weekly, monthly and flexible reporting needs

    Data-Driven Decision-making

    Built-in multiple query and analysis engines to help enterprises dig deeper into data values, observe business logic and insights, and make the best decisions through visual exploration and analysis of data, rules, models and cases.