What a true dark factory demands
A true dark factory is not an empty hall with cameras. It needs closed-loop authority so machines can start, stop, and recover without an operator beside the cell.
What lights-out actually means
Lights-out manufacturing (also called a dark factory) runs production with little or no routine human presence on the floor during normal operations. Robots, automated guided vehicles (AGVs) or autonomous mobile robots (AMRs), machine vision, and automated material handling execute the physical work. The decisive layer is autonomous control at the PLC level and in the higher orchestration systems: sequences start, stop, adjust parameters, and recover from many faults without an operator standing beside the machine.
Lighting, HVAC, and other human-centric infrastructure can be reduced or switched off because people are not required for routine shifts. Humans remain for scheduled maintenance, remote supervision, or genuine exceptions. Whole-plant unattended operation stays rare. Industry practice more often runs lights-sparse cells or overnight windows inside larger staffed halls. Why most floors still start from a people-heavy baseline is covered in most plants are still full of people.
Live visibility stops at observation
An operational digital twin works in a different layer. twinzo functions as a platform that ingests data from systems such as a real-time location system (RTLS), MES, ERP, WMS, and SCADA, then places the data in spatial context inside a live 3D model of the facility. Locations of forklifts, people, and materials appear in real time. Historical replay, spaghetti diagrams, and overlays give decision support so teams can see what is happening before they walk the floor. Medium-precision location around 1–3 m (3–10 ft) covers most aisle and buffer coordination. Exact floor position is a separate high-precision job.
The twin shows state. It does not write setpoints, issue start or stop commands, or take over the sequence logic that lives inside the PLCs and the manufacturing execution system. twinzo supplies visualization, spatial context, and decision support. Control authority stays where it already sits. The framing for that twin type is in operational digital twins show the shift.
Three equal observation jobs, still not control
Even as an observation layer, the twin has three equal jobs. Live internal logistics and shift coordination use current locations. The same view stores a spatially oriented dataset of place and time. Correlation and causality analytics ask which pairings keep preceding a stop. Those jobs support smarter staffed and lights-sparse operation. They do not by themselves create a dark factory. Product scope for the observation layer sits on internal logistics optimization, production monitoring, and the features overview.
Adjacent ideas that are easy to blur
Dark-factory automation and an operational digital twin both address the same physical plant, so the language slides together in conversation. A site already running automated cells still needs a clear picture of material flow, equipment status, and logistics movements. An operational digital twin can sit beside the existing control stack and give remote or central teams the live map they need for coordination and exception spotting. That adjacency is real.
It does not make the claims interchangeable. Live visibility improves the quality of human or higher-level software decisions. Autonomous control requires the write-back path that actually moves actuators. One layer observes and contextualizes. The other holds authority over the machines. Mixing the two in a single pitch raises expectations the visibility platform was never built to meet. Practical costs that still show up when a plant thins the shift, such as empty rounds and thin-shift balance, are covered in the empty-round problem that still sits in most plants and the thin-shift visibility gap.
A practical boundary check
1. Ask what writes to actuators - If the digital twin only displays PLC data, RTLS tracks, and MES metrics, it is delivering operational visibility. If a separate MOM or MES layer already orchestrates unattended runs, the twin can feed it richer spatial data and still stay the observation layer.
2. Separate lights-sparse cells from whole-plant dark claims - A cell that runs unattended overnight is not proof that the rest of the hall can go dark. High-volume, low-variation processes are the usual candidates. Mixed-model and exception-heavy work usually stay staffed.
3. Count empty-round and headcount cases as visibility returns, not dark-factory proof - Documented twinzo returns from claimable material pull and RTLS maps measure shorter search time and shared floor awareness. They do not measure closed-loop recovery without people. Example pages sit under seeing the floor when headcount drops and why large plants still pay for empty rounds.
Keeping that distinction clean lets a site run lights-sparse production in selected zones while using an operational digital twin across the rest of the floor without over-claiming what either system actually does.
Get in touch if you want to walk the observation-versus-control boundary on your own facility model.