One of the biggest improvements in my Home Assistant setup came when I stopped treating a single temperature reading as the truth for an entire room.
A sensor only knows the conditions exactly where it is mounted. Sunlight, closed blinds, an open door, airflow and the people in the room can all create a very different comfort level somewhere else. That is why one correct value can still lead to the wrong automation.
My setup uses Aqara temperature and humidity sensors as room references. I compare their readings with separate room sensors instead of trusting one value blindly. Presence then helps decide which room or reading should matter at that moment.
The logic is roughly:
Presence tells Home Assistant whether the room is actually being used.
Temperature and humidity describe what the room needs.
Door state shows whether the room is connected to the rest of the house or behaving as its own space.
Shading provides more context because direct sun and privacy can require different blind positions.
Home Assistant evaluates these signals together. This prevents one unusually warm or cold spot from making the whole climate system overreact.
I do not try to force every sensor to show exactly the same number. Small differences are normal. What matters is whether they tell the same story and whether the selected reference represents the place where people actually spend time.
This also makes the automation easier to tune. Instead of constantly changing the target temperature, I can change which reference is trusted under which condition.
How do you handle multiple temperature readings? Do you average all sensors, follow the occupied room, use the highest or lowest value, or rely on one fixed reference?