Enhancing U.S. Financial Compliance and Risk Management through Data-Driven Automation and Anomaly Detection

Abstract

The growing complexity of the U.S. financial ecosystem has heightened the demands of smart and data-driven compliance tools. These tools must detect emerging risks and identify fraud in various regulatory environments. The study proposes an integrated framework that automates anomaly detection and cross-silo data association. The goal is to improve financial compliance and risk management capacity. The analysis draws on several reproducible and publicly available datasets, such as the U.S. Small Business Administration (SBA), Paycheck Protection Program (PPP), loan information (2020 - 2021) and the Office of Foreign Assets Control (OFAC) Specially Designated Nationals (SDN) list (2021), the U.S. Securities and Exchange Commission (SEC) EDGAR Company Filings data (2021), the Home Credit Loan Default data (2021), and the ULB Credit Card. The suggested framework uses a combination of supervised and unsupervised learning models to assess behavioral deviations, financial irregularities, and entities that are related to sanctions. The study also provides integrity, privacy protection, and reproducibility of data through cryptographic hashing and tokenization. The Cross-silo linkage results indicate that regulatory, transactional and loan datasets improves the accuracy of fraud detection and false- positive alerts. This fusion exhibits hidden relationships between public funds, corporate filings and violation of sanctions that are often overlooked when datasets are analyzed separately. This study presents practical implications to enhance the transparency of audits and financial integrity in the country by mapping findings against the Anti-Money Laundering Act of 2020 (AMLA-2020) and the national financial priorities of FinCEN in 2021.