Pattern matching
Anomaly Detection groups your transactions into patterns and expects the tax results within each pattern to stay consistent.
How Anomaly Detection groups transactions
Anomaly Detection groups transactions into patterns based on 6 attributes:
- Transaction type
- Company role (buyer, seller, or middleman)
- Point of title transfer
- Origin (ship-from) address
- Destination (ship-to) address
- Product or commodity code
When all 6 attributes match across transactions, the system expects the tax rate, fee, and exemption amount to stay consistent for each jurisdiction. The system flags any deviation as a potential anomaly.
Data requirements
The system requires at least 30 calendar days of transaction history before it starts detecting anomalies for a customer. The system counts the 30 days from the date of the customer's first transaction. Days without activity count toward the 30.
By default, the system processes 12 months of historical data.