February 2022 Summaries

Joyjit Chowdhury, a senior data science engineer at Socure, combines a 15-year engineering background with ongoing work in machine learning, scalable data modeling, and real-time inference systems. Originally educated in India and now based in Phoenix, he emphasizes continual learning through courses and professional training, motivated by an interest in mathematics, data patterns, and the ways people and intelligent systems learn. His appreciation for linear algebra deepened as he connected concepts such as matrices, determinants, eigenvectors, and principal component analysis to their geometric meanings rather than memorized calculations. He also draws parallels between human cognition and artificial intelligence, using examples such as self-driving vehicles to illustrate the complexity of tasks humans perform intuitively. Inspired by mathematician Srinivasa Ramanujan, Chowdhury shares his learning through a Medium blog and seeks to help others develop a deeper understanding of technical ideas.

Socure describes how an institutional investment brokerage used its Phone RiskScore solution to detect and prevent pre-arranged trading fraud conducted by organized fraud rings. The scheme involved collusive securities trades intended to manipulate prices, followed by attempts to transfer illicit proceeds to another financial institution. By analyzing whether phone numbers were likely fraudulent and whether they were genuinely associated with the stated users, along with explanatory reason codes, the brokerage identified suspicious transfer destinations and blocked fund transfers. Socure positions its RiskScore products, which assess phone, email, and address signals, as tools used throughout the customer lifecycle to address application fraud, account takeover, marketplace fraud, and other financial crime risks.

The promotional piece argues that outdated identity verification systems can create customer friction, slow implementation, and provide limited value beyond initial onboarding, leading to lost revenue and growth. It presents Socure as an alternative, claiming a 95% conversion rate among users who begin verification, low-friction passive biometric checks, and automated machine-learning and computer-vision tools designed to reduce manual review and false positives. It also claims that its no-code mobile and web SDK integrations can be completed within days and that a configurable dashboard allows branding and workflow adjustments without coding. Beyond fraud prevention and compliance, the company frames identity verification as a means to improve customer engagement, accelerate onboarding, strengthen retention, and create competitive differentiation.

Aite-Novarica Group’s 2022 fraud trends report identifies sustained growth in application fraud and account takeover attacks, alongside emerging tactics such as PII tumbling driven by an expanding group of less technically sophisticated “citizen fraudsters” attracted by fraud opportunities during and after the pandemic. The report emphasizes that financial institutions must improve identity verification, authentication, onboarding efficiency, and the orchestration of increasingly diverse fraud signals while balancing security with customer experience. Socure positions its machine-learning-based identity fraud products and consortium data network as tools for detecting stolen, synthetic, and fraudulent identities, reducing false positives, and adapting models to evolving attack patterns through shared client feedback.

Gen Z, defined as Americans born between 1997 and 2012, is a diverse generation of nearly 90 million people entering adulthood and expected to become a major source of future consumer growth, with collective buying power already nearing $150 billion. Unlike millennials, many Gen Z consumers have limited credit histories, favor personalized, values-driven digital experiences, and often use alternatives such as buy now, pay later rather than traditional credit cards, creating challenges for businesses that rely on credit-header data and manual identity reviews. The text argues that these legacy onboarding methods can produce unnecessary rejections and customer abandonment, while organizations that improve identity verification and reduce friction can gain an advantage in acquiring thin-file and credit-invisible applicants. It presents Socure’s API-based identity platform, which uses machine learning, extensive data sources, and identity clustering to assess digital identities, as a solution, citing claimed improvements in approval, SSN-match, and address-match rates for young Gen Z users as well as revenue and account-acceptance gains for customers.

Knowledge-based authentication (KBA), which verifies users through answers to personal questions, was once widely used in financial services but is increasingly viewed as inconvenient, unreliable, and vulnerable to fraud enabled by compromised personal information. NIST stated in 2017 that KBA questions do not provide an acceptable secret for digital authentication, while industry research indicates that financial institutions are reducing their reliance on the method. The passage promotes Socure Sigma Identity Fraud as a passive alternative that assesses identity risk during onboarding, logins, profile updates, and high-value transactions, aiming to reduce customer friction while improving fraud detection and approval rates. In a comparison with a major investment bank’s legacy KBA system, Socure reported identifying fraud among the riskiest 16% of completed KBA quizzes, suggesting that most users could be verified in the background while higher-risk cases receive document verification.