A store operations assistant, in development
Zujaj is being built to turn camera observations into clear operational signals—so store teams can understand activity and decide where to act.
We’ll review your use case and camera setup.
Illustrative store footage
Start with a store question
Cameras capture activity. The useful question is what a store team can understand from it—and when the view is too incomplete to rely on.
A manager reviews whether another checkout is needed.
Store manager02Review staffing against observed demand.
Store manager03Investigate a pattern or change the camera view before interpreting a metric.
Store managerOne observation. A human decision.
Follow one fictional checkout event from a defined camera area to a manager’s decision. Every value in this example is illustrative.
Illustrative example · Observe
One fictional checkout area contains four people. The outline names the area being discussed; it is not model output.
A queue count is not a waiting-time estimate.
Concept walkthrough · not a live feed or released queue capability
Synthetic editorial scene · not customer footage
Observe · 14:32:00
One fictional checkout area contains four people. The outline names the area being discussed; it is not model output.
Understand · 14:32:10
The illustrative count crosses a configured threshold. Show the observation’s age and usable view before deciding what it means.
Act · 14:33:00
The proposed response is to review checkout coverage. Assignment and follow-up are planned—not a completed product action.
Review · 14:36:00
The later fictional view is partly obscured. Its count is unavailable; it cannot demonstrate improvement or the effect of a manager’s response.
A missing observation is unavailable, never zero.
Choose the first workflow to evaluate
Start with one store question. Camera suitability and measured coverage determine what a pilot can establish.
Compare observed entrance activity by hour only where the view is sufficiently covered.
The illustrative walkthrough shows four people against a configured three-person threshold.
An obscured area stays unavailable; it must not appear as an empty zone.
Synthetic editorial scene · not customer footage
Assess your existing setup
The proposed approach begins with a camera view and one useful question. Compatibility, lighting, occlusion, local compute, connectivity and support need review.
Start with your question, camera view and the decision it should inform.
Define signals, coverage, responsibilities and a success criterion.
Review authorized recordings and human observations within the agreed scope.
Measure what could be observed and what was missing before choosing the next step.
Intended data flow · evidence access is separate and authorization-dependent
Local processing is the intended design. Continuous video is not shown as a cloud upload.
Privacy and control
The intended design keeps continuous video on site while selected operational records synchronize. Evidence access depends on authorization and available retention.
Read privacy and deployment
Recording files, reference images and review material stay local. Configuration and pilot metadata can reach the workspace; these are not accepted analytical results.
Before we begin
We assess your existing camera or recorder, its view and access, lighting, local compute and connectivity. Compatibility is specific to the setup; it is not universal.
No. The current workspace supports configuration and recorded-video pilot work. Entrance counting remains under evaluation. Checkout and zone analytics and action workflows are planned; the walkthrough is illustrative.
Only what the view and evaluated capability support. Field of view, lighting, overlap and occlusion affect usable coverage. A count alone does not establish waiting time, intent or sales conversion.
Local processing is the intended design, and current recording files remain local. Selected configuration and pilot metadata can reach the workspace. Evidence access is separate and permission-dependent.
Observation freshness and coverage can be affected. Local event buffering and later synchronization are intended architecture, not a verified offline promise. We assess the operating limits for each pilot.
One store question, permission to use the relevant camera view and a person who can support evaluation. We agree scope, responsibilities, data handling and a success definition before starting.
No public pilot price or pricing tier is confirmed. Store scope, camera suitability, processing needs and support requirements need review before a specific proposal can be made.
A focused first step
Tell us about your stores and camera setup. We’ll review whether a focused Zujaj pilot is a fit and agree what it should demonstrate.
Request a pilotFit review → agreed scope → evaluation plan
