Let’s treat this as something we can test rather than something we guess at.

What we know. Full charge around 6 p.m. Dead around 2 a.m. That is roughly eight hours, down from what used to be twelve or more. We do not yet know whether the change was sudden or gradual, or whether anything else changed at the same time: a software update, a new app, a new game, a new place where the phone sits overnight.

Hypotheses.

H1: An app is running more than it used to, especially in the evening and overnight.
H2: The screen is on for more of that window than it used to be.
H3: The phone is spending energy searching for a weak signal wherever it spends the night.
H4: The battery itself has lost capacity.

Experiments, cheapest first.

Battery health. Read the battery health figure in settings. If it is well under 90 percent, H4 is likely and the rest matter less.
Usage log. Tomorrow at 2 a.m., or first thing in the morning, open the per-app battery breakdown and note the top three apps and the total screen-on time since 6 p.m. A large screen-on number supports H2. One app at the top supports H1.
Controlled night. Charge to 100 percent at 6 p.m., then leave the phone face down in airplane mode in the room where it usually spends the night. If it is still mostly charged at 2 a.m., the drain comes from use or signal, not the battery.

Run the first one now. It takes thirty seconds and may end the investigation.

Programming AI with Personas