Why Voids and Discounts Are the Primary Theft Vectors
Transactional theft in service businesses most commonly operates through voids (reversing paid transactions and pocketing the cash), unauthorized comps (giving free service in exchange for cash payments outside the system), and discount abuse (applying unauthorized discounts to friends or family). All three leave traces in transaction data. Ezra finds them.
Pattern Detection vs. Individual Flag Review
Reviewing individual void transactions is inefficient and ineffective. A manager who voids three transactions on a busy Saturday is probably correcting errors. A manager whose void rate is 4x the network average across 30 shifts is a different signal. Ezra's anomaly detection operates on patterns, not individual transactions—surfacing the meaningful signals without burying operators in noise.
Six Detection Surfaces for Transactional Anomalies
Ezra monitors voids, manager overrides, comps, discount percentages, cash variance, and productivity anomalies. Each surface is tracked against location-specific baselines with configurable thresholds. Anomalies are surfaced as a triaged feed—the highest-risk flags first—with direct links to the source POS record for investigation.
From Flag to Investigation Without Spreadsheets
Traditional exception reporting requires pulling a report, exporting it, filtering it, and then locating the relevant transactions in the POS for investigation. Ezra shortcuts this entire process: each flag links directly to the source POS record. Investigation starts the moment the flag appears.
Configurable Thresholds Reduce Alert Fatigue
Not every void is theft. Ezra's thresholds are operator-validated and configurable per location—so the void rate that triggers a flag at a low-volume boutique location is different from the threshold at a high-volume flagship. Alert fatigue is the enemy of operator trust; meaningful thresholds are the solution.