Fraud Detection

Protect your organizations from frauds with advanced AI.

Real-Time Transaction Fraud Detection

Working with one of the largest banks in Asia, Graphen developed a real-time transaction fraud detection system to integrate customer data and transaction information, effectively identify the risk of fraud in customer transactions through aspect management.The technology ensures suspicious transactions are alerted before the transaction execution. The final decision will be made by the business department through the same system UI.

It provides an aggregate risk assessment to predict cybersecurity risks of all entities within the organization.


Anti-Money Laundering

Graphen developed an AML system that imports behavior models to detect money laundering patterns, and utilizes machine learning management module for model optimization.

Comprehensive detection is carried out by four major risks including Entity Risk, Known Suspicious Patterns, Relationship Risk and Behavioral Risk. AI calculates a comprehensive risk score based on the four major risks to facilitate anti-money laundering monitoring personnel to prioritize high-risk cases. Real-time transaction monitoring catches money laundering through rule review and money laundering list scanning, it then sends notification through system connection.

Taking into consideration of complex relationships and cashflow between entities and organizations, Graphen's AML solution has proven to be more efficient in detecting suspicious activities with lower false alarm rates and higher accuracy.


Insurance Fraud Detection

Based on the advanced graph database, the Insurance Fraud Detection solution performs graph analysis to find out hidden relationships and uses machine learning to create behavioral models through known suspicious patterns. These two components complement each other to optimize results and detect insurance fraud. Graph indicators for graph analysis can also be used as features into the machine learning model. Meanwhile, the machine learning feature value ranking list and blacklist label, can enter the graph database to create a new graph, through which the results can be further optimized.

By combining these advanced technologies together, Graphen provides a secure environment to allow corporations store their sensitive data in the Cloud. The secured data can be utilized by AI applications, but the organization still has full control of the data flow.

Graphen Insurance Fraud Detection can help detect: medical insurance fraud, property insurance fraud, accident insurance fraud, etc.


Financial Agent Fraud Detection

Graphen's Financial Agent Fraud Detetion adapts artificial intelligence and graph database to detect potential fraud of financial agents and protect banks from the inside. It establishes a relationship graph of behavior pattern detecting violations and frauds of any agents. It then builds a fraud map model where the system conducts pattern comparisons based on agents' known behavior and detect abnormal behaviors among financial agents.

The solution implifies the rule management process through a visual interface, adjusts existing rule thresholds, parameters, pre-values, etc., at any time, and supports version control mode at the same time. It designs a user interface and adds new rules as needed.


Request a demo today.