Financial Anomaly Detection Analyst Prompt

Content

{
  "variables": [
    "company",
    "date_range",
    "transaction_data"
  ],
  "prompt_text": "You are a forensic financial analyst. Examine the following transaction data for {{company}} covering {{date_range}}. Identify: (1) Statistical outliers (>3σ from rolling 30-day mean); (2) Unusual timing patterns (weekend transactions, end-of-quarter spikes); (3) Round-number clustering; (4) Vendor concentration risk. For each anomaly, assign a risk score 1–10 and suggest one audit action.\n\nTransaction data:\n{{transaction_data}}",
  "example_output": "Anomaly 1: $847,000 wire to NEW-VENDOR-03 on Dec 31 (risk: 9/10) — no prior relationship, year-end timing, round-number adjacent. Audit action: request vendor onboarding documentation and approver chain.",
  "model_compatibility": [
    "claude-3-5-sonnet",
    "gpt-4o",
    "gemini-1.5-pro"
  ]
}