A break-in at night, a crowd by day.
One model reads both, on the same camera, without being told which one to expect.
lost by Italian retailers to theft and shrinkage in 2024, around 1.2% of annual revenue.
Barometro dei Furti nel Retail in Italia 2024, Checkpoint Systems with NielsenIQ
of the loss is theft by people in the shop
is internal
is supplier or administrative error
Shrinkage is the number a retailer can see in the accounts at the end of the year. What nobody has is the moment it happened, on which camera, and what it looked like.
Same cameras. A different answer.
How it is handled today
Hundreds of alarms on movement, to open and read.
At three in the morning the camera watches for a door being forced; at six in the evening, a room full of people. A rule system needs two installations.
With Focoos
One alarm, already analysed.
One model reads both: a door being forced at three in the morning, a room filling at six in the evening. The night list and the day list are the same twenty, ticked differently.
What Focoos wrote about this clip
A person wearing a hoodie and carrying a bag approaches a glass door, uses an object to break the glass, and then walks away with the bag.
Real surveillance footage, and the model’s own sentence about it. The event types are the ones it raised.
What it watches for on this kind of site.
The usual list for this point of view. It is yours to change, camera by camera: the model is the same everywhere, and only the ticked boxes differ.
These are event types from the same twenty. Nothing is built for this market, which is why a site with a forecourt, a shop and a yard runs one model and not three systems.
On the cameras you already have.
The same cameras, the same recorder, the same cabling. What changes is that the alarm arrives with a sentence instead of a timestamp.
The brain: observes, understands and decides on-prem, on your own network.
Where your team settles the events, from every site you run.