For retailers
Catch theft as it happens.
Shoplifting gestures spotted in the act, repeat offenders recognised at the door, and the whole store understood — from the cameras you already own.
Loss prevention
It knows what concealment looks like.
AEY AI watches for the gestures that precede theft and alerts staff while the shopper is still in the aisle — early enough for customer service, not confrontation.
Bag concealment
An item goes into a backpack, handbag or carrier instead of the basket — staff know in seconds.
Clothing concealment
Items hidden in coats, jackets, trousers or waistbands, flagged mid-gesture.
Pram & stroller concealment
The baby compartment, the under-basket and hanging bags are watched too.
Eating in store
Products consumed in the aisle raise an alert before the empty wrapper hits the shelf.
Suspicious behaviour
Frequent glances at staff or cameras, a hand held in an open bag, lingering with merchandise and no basket.
Your own rules, in plain English
Type what to watch for — “alert me when someone puts an item in a bag” — and AEY AI enforces it. No model training, no rule builder.
How it works
Say it. It watches. Staff step in.
- 1
Say it
Pick a ready-made command or write your own: “a person concealing a store item in a bag or coat.”
- 2
AEY AI watches
A vision model reviews every detection frame against your command — on the box, so footage never leaves the store.
- 3
Staff step in early
A discreet alert with a clip. An offer of help at the shelf beats a stop at the door.
Know the face at the door.
Faces are read on the box and gathered into people across your estate. Flag a known shoplifter once — the next visit raises an alarm as they walk in.
Recognised before they reach the shelves.
Facial recognition built for retail loss prevention, not surveillance theatre.
- Watchlist alarms the moment a flagged person returns
- Search by photo — see everywhere a face has been across every store
- One person, one history, however many sites they visit
Built for a DPIA, not around one.
Recognition stays off until an admin turns it on, runs on your own box, and forgets on your retention schedule.
- Read on the box — no cloud vision API, no fee per face
- Every flag signed and audited; a human confirms each match before anyone acts
- Delete means deleted — signatures cleared from the index, on schedule
Beyond security
The same cameras, running the store.
Loss prevention pays for it. These come with it.
Heatmaps — where shoppers stop, pass, and dwell.
Line counting — entries, exits, and conversion by daypart.
Time in zone — which displays hold attention, and for how long.
Queue alerts — know the till is backing up before customers walk.
Footfall against transactions, on one chart.
Questions
Straight answers.
Does it work with our existing CCTV?
Yes. AEY runs on the cameras and VMS you already have — no rip-and-replace. An AI Box on your network does the seeing, and footage stays in the building.
Is shoplifting detection accurate enough to act on?
It is accurate enough to put the right clip in front of a person, fast. Every alert carries the frame and a description, and your policy decides what staff do next — most stores lead with an offer of help, which is usually enough.
Is facial recognition legal for retail?
Used properly, yes in the UK — with signage, a DPIA, and a human confirming every match. The product asks for all three before it will run, recognition is off until an admin enables it per camera, and biometric data expires on your retention window.
How do staff get the alerts?
Push to iOS and Android, on-screen in the dashboard, or into your existing alarm flow — including your monitoring provider, if you use one.
Can it learn our own products or store layout?
Yes. Zones, lines and commands are per camera, and you can train custom models on your own footage — no data-science team.