Financial fraud networks are increasingly organized and industrialized, with criminals disrupting the financial ecosystem through high-frequency malicious complaints. What appears to be ordinary complaint leads may conceal a criminal chain of illegal activities. In May this year, Fenqile, a subsidiary of LexinFintech Holdings Ltd. (NASDAQ: LX), publicly called for information on financial fraud crimes, sparking widespread attention. To date, over 5,300 individuals have submitted more than 17,000 tip-offs, with several leads assisting police in initiating investigations. Confronted with vast and complex data, Fenqile has deployed a fraud-identification intelligent agent. Leveraging artificial intelligence, this system performs information extraction, risk identification, and relationship mining, upgrading the process from clue discovery to case analysis and risk management from reliance on manual expertise to technology-driven detection.
Previously, handling fraud-related complaint cases depended heavily on manual judgment. Staff had to painstakingly analyze large volumes of complaint texts, user feedback, and historical data, a process that was not only time-consuming but also struggled to quickly uncover the syndicated patterns hidden within complex information. Now, with artificial intelligence, Fenqile is transforming this process. The company's fraud-identification intelligent agent applies AI technology across the entire fraud management lifecycle. This agent can comprehensively analyze multi-dimensional data, including customer service records, post-loan data, user profiles, and clustering correlations. Using a locally deployed large language model, it automatically extracts information, matches modus operandi characteristics, and uncovers relationships to identify potential risk leads. In the case analysis phase, the system helps identify connections between cases, detects abnormal patterns that manual screening might miss, and generates analytical reports to support decision-making for subsequent case handling.
By integrating AI capabilities, the fraud management model has shifted from manual experience to data-driven and intelligent identification. Real-world application data shows that since the fraud-identification intelligent agent went live, the monitoring and analysis process has become automated, reducing the case analysis cycle from approximately 0.5 days to just a few minutes—a 90% efficiency improvement. Additionally, the agent continuously learns from massive complaint data and extracts features, enabling it to uncover abnormal patterns and relational characteristics in fraud operations. In practice, the number of newly identified abnormal features has steadily increased. Compared to manual screening, the accuracy of fraud identification has improved more than tenfold, effectively helping the management team quickly pinpoint high-risk cases from a large pool of leads.
In combating financial fraud, relying solely on technology is insufficient; the key is forming a closed loop from lead discovery to enforcement. To achieve this, Fenqile has established a governance mechanism that combines AI-driven intelligent identification with professional human judgment. The AI agent handles large-scale information processing, risk screening, and relationship analysis, boosting case detection efficiency. Meanwhile, the professional governance team conducts manual reviews, evidence collation, and follow-up actions for high-risk cases, ensuring more accurate risk identification and more effective case progression. Fenqile's public notice soliciting tips on illegal financial fraud activities (partial) is part of this effort. With the participation of regulators, judicial authorities, institutions, and the public, the collective force against financial fraud is steadily growing. Over the past two years, Fenqile has assisted police in Shanghai, Fujian, Shandong, Jiangsu, Liaoning, and other regions in solving nearly 100 related cases, successfully dismantling 26 professional criminal gangs, with over 100 individuals facing legal penalties.
As a leading new consumption digital technology service provider in China, LexinFintech Holdings Ltd. is committed to the mission of "making consumption better through technology." It continuously creates new consumption methods, leveraging technology to enhance the efficiency and operational capabilities of over 20 million merchants and nearly 200 financial institutions, serving more than 250 million high-quality, high-growth consumers in China. LexinFintech Holdings Ltd. is accelerating its AI deployment, using artificial intelligence as a key driver of new productive forces. It has developed an AI agent system with financial adaptive capabilities, built an "AI composite agent matrix," and deployed over 100 AI agent positions covering core business scenarios such as risk control, customer service, operations, and research and development, truly making AI a productive force for business operations. LexinFintech Holdings Ltd. will continue to refine its intelligent risk management system, enhancing the application of AI, big data, and other technologies in fraud identification, risk prevention, and consumer protection, using technological power to support the healthy development of the financial ecosystem.
Comments