Every controller in the office shows 23 °C. Why are some people hot and others cold?

The operator sets every controller to 23 °C. An hour later, one office asks for more heat and another for less. After several manual adjustments, the same thing happens again.

I would not start by looking for the ‘right temperature for the whole floor’. First, we need to establish exactly what each unit measures and controls.

A setpoint is a target, not a measurement

The figure of 23 °C is the target value for the control algorithm. It does not prove that this is the temperature at every workstation.

The sensor may be in the indoor unit or at another designated control point. Readings near the ceiling or air intake may differ from the average temperature in the occupied zone. Daikin explains this effect for its systems: the result depends on the unit’s location, the room and the fan mode. ¹

The first practical step is therefore to compare the control sensor reading with an independent measurement at a representative location, rather than immediately applying an offset of several degrees.

Two identical offices can have different loads

For example, one office is sunlit while the next is shaded. Six people work in the first and one in the second. The setpoints are identical, but the cooling or heating demand differs.

If the system only supports a common operating mode within one refrigerant circuit, local controllers will not provide fully independent heating and cooling. For Mitsubishi Electric, this is a fundamental distinction between the Y heat pump configuration and R2 with heat recovery. ²

It is therefore important to check how rooms are grouped by outdoor unit. Physical zoning matters more than the number of controllers on the walls.

Comfort is about more than temperature

A strong cold air jet near a workstation and poor mixing elsewhere in the room can create different sensations even when thermometer readings are similar. Supply air direction, fan speed, unit location and obstructions in the airflow path need to be assessed together. ³

In this situation, changing the setpoint may help one person while making conditions worse for another. Often, the right question is not ‘what temperature should we set?’ but ‘how can we supply air without causing discomfort?’

Check more than just the VRF system

Imagine a room where the VRF system is cooling while the ventilation supplies excessively warm air without a justified need. Or a radiator continues heating the room even though cooling is already active.

These are scenarios to investigate, not a diagnosis based on a single screenshot. We would compare temperatures, commands and the actual operation of associated systems. Checking the implemented control algorithms and their compliance with the design is part of commissioning, not a sign that equipment must be replaced immediately. ⁴

How to make the problem visible

A log is enough to start with: zone, setpoint, actual temperature, mode, time and reason for intervention. It is useful to record which indoor and outdoor units serve each zone.

This turns a general complaint that ‘people are uncomfortable’ into a specific picture. For example, the problem occurs only in the afternoon on one façade, or only after the shared circuit changes mode.

AI analytics should be applied to this picture, not to a set of identical setpoints. First, understand the cause. Then check which solution genuinely improves conditions.

Good office control does not mean identical numbers on every controller. It means agreed comfort conditions in every zone.