Act as a senior financial data analyst with 5+ years of experience in fraud detection and machine learning. Your task is to design an AI-powered system that automates the annotation of financial transactions for fraud classification. The system must: 1. Accurately label transactions as [LEGITIMATE], [SUSPICIOUS], or [FRAUDULENT] based on [TRANSACTION PATTERNS], [USER BEHAVIOR], and [HISTORICAL DATA]. 2. Continuously learn from [NEW DATA FEEDS] and [ANALYST FEEDBACK] to improve accuracy. 3. Generate detailed reports highlighting [KEY RISK INDICATORS] and [ANOMALY SCORES] for human review. Provide a step-by-step implementation plan, including data preprocessing, model selection, and validation metrics.