The current pilot combines local person and table occupancy detection, staff table checks and staff-recorded Ready status. These signals are under field validation in Mumbai.

Existing CCTV may be suitable. Camera access, recorder compatibility, viewing angles and table coverage are assessed before setup.

Camera inference runs locally on suitable venue hardware. CPU, memory and competing workloads need assessment; compatibility with a billing PC or tablet is not guaranteed.

Tracks observed staff visits and time since a table check. Staff record Ready status separately; a visit does not prove a greeting or food delivery.
Sangati™ tracks table occupancy and observed staff checks. These observations can show the time since a check, but cannot establish what was said, whether food arrived or whether a guest requested the bill.
Ready status is recorded by staff. Detection and timing are being validated against floor observations; kitchen bottlenecks, inventory savings and revenue effects are not established pilot outcomes.
Built by an operator who ran the floor — not a software tourist. We'll assess honestly whether your camera coverage is a fit.
Existing CCTV may be suitable. Camera access, recorder compatibility, table coverage and device resources must be assessed before setup. A billing PC or tablet is not automatically suitable.
No. Sangati uses person detection and table observations, without facial recognition, identity inference or guest profiling.
Person and table occupancy, staff table checks and staff-recorded Ready status are under field validation. A staff visit does not prove a greeting or food delivery. Bill intent and kitchen bottlenecks are not validated capabilities.
No. Camera inference runs locally. The operational dashboard can be accessed remotely by authorized users, so operational records may leave the venue. Dashboard access, retention and any diagnostic access need to be agreed for each venue.