A sensor measures a condition. A platform connects that measurement to an asset, a customer and a process. But an operational decision often requires something more. Knowing that a container has reached a certain fill level, that its temperature has increased or that a vehicle has changed location does not automatically mean knowing what to do. The data must be interpreted in its context.
Different data, different levels of reliability
GPS location, temperature, tilt, fill level and images do not always have the same weight. A value may be recent but inconsistent with other data. An image may be partially obstructed. A threshold may indicate a real anomaly or a condition that is part of the process.
For this reason, Twin Control does not treat every measurement as an isolated truth. Information is connected to the asset, its history and the activities currently in progress, providing a more complete view of the situation.
The result is more effective operational control: less time spent interpreting separate data and more attention focused on events that genuinely require action.
Understanding why something requires attention
An alert is useful only if it helps people make a decision. For this reason, an alert should show which data generated it, how recent the data is and which rule was applied. This allows the operator to understand the reason for the priority, distinguish a real problem from a false alarm more easily and explain the reason for the action taken.
This is the principle behind Explainable AI (XAI): not simply identifying what requires attention, but making the reasoning behind that indication understandable.
Automation where it helps, human control where it matters
Not every activity requires the same level of supervision. Repetitive, predictable and easily reversible operations can be automated. When a decision may affect customers, costs, safety or compliance, however, more explicit human control is needed.
Twin Control therefore adopts a Human-in-the-Loop approach, meaning human supervision within the decision-making process. The concept describes a model in which an automated or AI-based system analyses data, detects anomalies, prioritises situations or suggests actions, but does not operate entirely on its own. A person remains involved in the steps that require assessment, confirmation or responsibility.
In Twin Control, for example, the platform can detect an unusual fill level, compare it with other data and suggest that a container requires attention. The operator can then verify the information, understand why the situation was flagged and decide whether action is needed.
The principle is therefore simple: the machine supports the decision; the person retains judgement and final responsibility; the platform collects data, highlights anomalies, prioritises situations and supports analysis. The final assessment, however, remains with people. The goal isn’t to replace the operator’s experience, but to enable them to make decisions more quickly and with better information.
Decision intelligence also means responsibility
A Decision Intelligence platform should not make the decision-making process invisible. On the contrary, it should make it possible to understand which information was considered, what level of reliability it has and why a particular situation was flagged. In this way, automation and responsibility can work together.
Technology collects and organises a volume of data that would be difficult to monitor manually. People retain judgement, the ability to interpret context and responsibility for decisions.
It is this balance that transforms monitoring into genuine operational support.