Finance & Insurance

Business Level - Artificial Intelligence for IT Operations (AIOps)

  • AIOps utilizes the massive data generated by the daily operation of the internal IT system of a general enterprise to quickly detect and predict system abnormal events/behaviors, and provide instant alarms.
  • In a dynamic IT system environment, abnormal events/behaviors will change over time and in different scenarios.
  • Financial industry information systems are facing challenges: (1) Information personnel with different expertise in network, server, and application programs are required, and they need to integrate system environments customized by different manufacturers and monitor information; (2) Distributed architecture operating environment, diverse Sexual application services.
  • The cloud environment of the financial industry is facing challenges: (1) Massive monitoring indicators; (2) New and old versions of the system, there are old and new versions of monitoring indicators (increased difficulty).
  • With AIOps: (1) Improve system service quality and enhance system stability; (2) Advanced current maintenance and operation management, visual monitoring, easy to find key indicators that cause system abnormalities, quickly locate abnormal problems, and reduce maintenance time.

    • Business level - credit card

  • In recent years, the national online shopping demand has grown rapidly, and the demand for online shopping during the epidemic was also one of the reasons, resulting in a substantial increase in the amount of credit card fraud. According to the data of the National Credit Card Center of R.O.C, the amount of stolen credit cards in China in 2021 will be 1.96 billion, and the accumulated amount of stolen credit cards from January to September 2022 will reach 1.583 billion, and it is expected to exceed 2 billion by the end of the 2022. Taking 2021 as an example, if the detection rate of credit card theft is increased by 1%, the loss of 19.6 million yuan can be reduced.
  • Real-time detection of fraudulent credit card behavior can greatly reduce the cost and loss of financial industry investigations, and also improve customer confidence and satisfaction due to real-time detection of fraudulent credit card behavior.
  • Through the 1H3A abnormal behavior detection method, a credit card network transaction fraud (fraud) detection model is built on the original early warning system, which can achieve enhanced fraud (fraud) transaction judgment information, improve risk control, and reduce risks.

    • Business level - property insurance

  • According to the research of the American Insurance Information Institute, the amount of property insurance fraud every year accounts for about 10% of the annual insurance claims. Taking the United States as an example, the estimated annual loss of insurance crimes in the United States is as high as 80 billion US dollars, accounting for about 10% of the amount of claims. In 2019, the number of insurance crimes in Taiwan is estimated to be as high as NTD 10.5 billion. If the defrauded amount is reduced by 1%, the loss of NTD 10 million can be reduced.
  • Claims fraud cases require early detection and investigation of abnormalities during the claims process.
  • Through the 1H3A abnormal behavior detection method, a fraud detection model is built on the original early warning system, which can be gradually applied to the compensation case processing mechanism, strengthens the judgment information of fraudulent transactions, improves risk control, and reduces risks.

    • Business Level - Property Insurance (Claims Fraud Applied to Automobile Insurance)

  • Import the car accident hotspot map, and judge the car accident as a fraud possibility input model based on key information such as the location and time of the car accident.
  • The model is combined with an open map, and through the latitude and longitude of the car accident, the information of the corresponding road section is obtained to strengthen the judgment ability of the model.

    • Cooperating with us, possible fraudulent transactions and fraudulent transactions can be prevented in advance.

    View application cases