Money laundering—the concealment and integration of illicit proceeds into the formal financial system—undermines the trust and fairness of global financial systems, presenting enormous challenges to investors, regulators, and commercial enterprises. Traditional detection methods based on rigid rule-based systems and manual auditing have proven insufficient in combating increasingly sophisticated laundering schemes. This paper demonstrates how data science, commercial domain knowledge, and machine learning—specifically, decision tree models—can be synthesized to enhance real-time detection of suspicious financial activities. Through a comprehensive workflow involving synthetic transaction data generation, exploratory data analysis, and predictive modeling, critical patterns such as transaction amount, timing, customer risk profiles, and transaction type emerge as powerful indicators of money laundering behavior. Bar diagrams and visual analytics visually support the findings, illustrating feature importance rankings and identifying high-risk transaction segments. The commercial impact of this approach includes proactive regulatory compliance, significant workload reduction for compliance analysts, and minimal customer friction through reduced false positives. This research highlights how student-level expertise combined with interpretable AI tools can effectively bridge the gap between traditional commerce education and modern financial technology compliance solutions. The decision tree model achieved 99.93% testing accuracy with a precision and recall of 99.82% each, demonstrating the viability of automated AML detection systems in real-world banking environments.