FraudEx
Explainable fraud detection that shows why a transaction was flagged.
- 01
Challenge
Auditing financial records is manual and reactive. Suspicious rows hide inside thousands of transactions, and most tools flag them without explaining why.
- 02
Approach
A Next.js frontend over a Python FastAPI backend running Benford's Law, Z-score and IQR outliers, vendor concentration, round-number and duplicate checks. Every flag carries its reason.
- 03
Outcome
Upload a CSV and get a risk score, flagged transactions, a Benford's Law chart, and a plain explanation in under two minutes. It informs auditors rather than accusing.