4A LABS AI
storefrontCVAI STORE VISUAL ANALYSE

Your cameras are already there.
Nobody is
watching them.

CVAI Store Visual Analyse connects to your existing IP cameras and keeps checking the rules you define. No staff behind the counter, a queue getting too long, a store that did not open on time: each one arrives as an instant alert with an evidence frame.

START A PROJECTarrow_forwardHOW IT WORKS
videocam Works with your existing IP camerasnotifications_paused No false alarmslock Video never leaves the store
Overhead view of a coffee shop: staff behind the counter, customers in the waiting area and a person at the door, each framed head to toe; zones drawn on the floor.
groups

STAFFING

Is anyone behind the counter?

people_alt

QUEUE

An alert the moment the line is too long

schedule

OPENING

Real opening and closing time

photo_camera

EVIDENCE

Every alert comes with a frame

Two baristas seen from behind, each framed by a box, inside a zone drawn on the floor behind the counter.

Representative scene.

RULE 1

Is There Staff Behind the Counter?

An empty counter at peak hour is a lost sale. You set how many people must be behind the counter for each day and hour; when the count drops below it and stays there, you get an alert.

  • scheduleA different minimum per day and time slot
  • accessibility_newBending down or leaving the frame for a moment is not an alarm
  • photo_cameraThe evidence frame is saved with the alert
  • groupsCounts people, never identifies them
Five customers standing in line in front of the counter, each framed; the waiting zone is outlined in red.

Representative scene.

RULE 2

When Did the Queue Get Too Long?

You usually hear about a long queue after the customer has left. People waiting in front of the counter are counted continuously and the branch is alerted the moment your threshold is passed.

  • people_altAn alert the moment the threshold is passed
  • directions_walkPeople walking through are not counted as queue
  • monitoringBreaches charted through the day; peak hours become clear
  • event_availableStaff planning based on real data
A person walking in through a glass entrance door; a line with arrows in both directions is drawn across the threshold.

Representative scene.

RULE 3

Did the Store Really Open on Time?

Today the answer is whatever staff report. A direction-aware line at the door counts every pass; the first and last pass of the day give the real opening and closing time.

  • swap_horizEntries and exits counted separately
  • scheduleActual opening and closing time, no self-reporting
  • reportLate openings and early closings reported automatically
  • compare_arrowsBranches compared on objective data
HOW IT WORKS

Installed on Top of Your Existing Cameras

01

Connects

Connects to your existing IP cameras. No new cameras, no new cabling.

02

Sees

Detects people and tracks them frame by frame. No identification: it only knows “a person is here”.

03

Measures

Counts against the zones and lines you draw in the browser: how many, how long, in or out.

04

Verifies

A momentary situation does not raise an alarm. The condition has to persist; the same alert is not repeated.

05

Notifies

Once a violation is confirmed, an alert is raised, the evidence frame is saved and it lands on the panel.

notifications_paused

It does not cry wolf.

A manager who gets thirty alerts a day stops looking at any of them. An alert is raised only when a condition persists; short passes are ignored and the same alert is not repeated during a cool-down period.

START A PILOTarrow_forward
FEATURES

Rules Are Defined From the Panel, No Code Needed

draw

Zone Drawing in the Browser

Draw zones, lines and areas over the camera image with your mouse.

rule

Rule Panel

Camera, zone, rule type, days and hours, threshold: all from the panel.

timer

Dwell Time

Measures how long people stay in a zone you define.

swap_horiz

Entry / Exit Counting

Direction-aware visitor counting; entries and exits never mix.

thermostat

Heatmap

Shows where in the store the density builds up.

blur_on

Privacy Masking

Selected areas of the image can be masked to meet data protection rules.

upload_file

Camera Inventory

Cameras are imported in bulk from Excel or CSV; recorders are found by scanning.

movie

Test With a Video File

Try a rule on recorded video before connecting to a live camera.

photo_camera

Evidence Frame

Every alert keeps the image of that moment.

PRIVACY BY DESIGN

memory Analysis runs on a small device inside the storeface_retouching_off No face recognition, no biometricslock Only counts and events leave the store, never raw video
FAQ

Frequently Asked Questions

Does CVAI Store Visual Analyse use face recognition?

expand_more

No. It only detects that a person is present and counts people in the zones you draw. There is no face recognition or biometrics, and only counts and events leave the store — never raw video.

Do we need new cameras?

expand_more

No. It connects to the IP cameras you already have, with no new cameras or cabling.

How does it avoid false alarms?

expand_more

An alert is raised only when a condition persists for a set time; short passes are ignored and the same alert is not repeated during a cool-down period.

Start with one store and one rule.
We will show you the rest.

We connect to the cameras you already have and start with a pilot in a single branch.

  • check_circle The first call is free and technical
  • check_circle We start under NDA
  • check_circle Scope and price are clear up front

PROJECT REQUEST FORM