Theft rarely looks like theft on a monitor — the person browses normally right up until the item is concealed. The theft module is built to catch that concealment moment itself, not just flag it after an inventory count comes up short, and to hand loss prevention a timestamped clip instead of a shrinkage number with no evidence behind it.
Theft Detection is currently finishing testing ahead of general rollout. Early-access pilots are open now — talk to us about running it at your site.
Concealment under clothing or into a bag, unauthorised removal from shelves, displays or storage, unusual lingering near high-shrink SKUs, and repeat visits with matching behaviour patterns. Target confidence for the module is 93%+ — high enough to route an alert to floor staff in real time instead of a note for the end-of-day audit.
Retail floors, supermarkets, warehouse and yard operations, click-and-collect counters, showrooms, and electronics stores — anywhere shrinkage shows up on a P&L and nobody has time to review a week of footage to find out where it's coming from.
A theft caught after the fact is a write-off and a line in a shrinkage report. A theft flagged the moment concealment happens is a real-time intervention, a timestamped clip for the case file, and one less item in the loss-prevention team's end-of-week reconciliation — the difference between a police report or insurance claim with evidence attached and one without.
Theft Detection + Face Recognition — the combination doesn't just catch the incident, it flags the person again the next time they walk in, so loss prevention isn't starting from zero on every repeat visit.
