The Multi-Location Detection Problem
Each additional location multiplies the transaction volume—and the surface area for undetected anomalies. A 10-location franchise might process 10,000 transactions per day. No manual review process catches the meaningful patterns in that volume. AI anomaly detection does.
Network-Level vs. Location-Level Anomalies
Some anomalies are visible at the location level: a specific location's void rate is running above its own historical norm. Others are only visible at the network level: a location's void rate looks normal on its own but is an outlier relative to comparable locations in the portfolio. Ezra detects both simultaneously.
Six Detection Surfaces
Ezra Loss Prevention monitors six detection surfaces across every location: transaction voids, manager overrides and comps, discount percentages, cash variance, productivity anomalies, and cross-location behavioral comparison. Every flag links to the source POS record for immediate investigation.
Triaged Feed, Not an Exception Spreadsheet
Traditional loss prevention delivers a 200-line exception report. Ezra delivers a triaged feed—the highest-risk anomalies first, with context and source data immediately accessible. The operator spends their investigation time on the signals that matter, not filtering through noise.
Scales Automatically as You Add Locations
As operators add locations to their portfolio, Ezra adds them to the detection layer automatically. There is no re-implementation, no manual configuration per location—just expanded coverage as the network grows.