fwiw this just depends a lot on the OEM. All OEMs have a non-trivial mapping of assumed energy content to shown percentage, in part because that's just not trivial to do, in part psychology. (The same is true for fuel gauges). Some brands manipulate the consumption numbers a lot (iirc - Tesla fudges the numbers the most e.g. they don't count energy consumption when not moving).

For most the consumption of an EV is shown rather accurately and is really just (energy drawn from the HV battery as measured by the BMS, which is generally very accurate for safety reasons) divided by (kilometers driven in the current interval).

As far as I know nobody includes charging losses in the readout. Personally I think this is correct for range estimation purposes. If you're swinging at a cost perspective, this of course means you're going to always miss around 3-20% depending on charging method, temperatures, car model etc. (there are fairly significant differences in OBC efficiency across models and also across AC power, low power charging gets inefficient quickly, while 7-11 kW often gets you north of 90%. DCFC is usually more efficient accounting-wise, because fast chargers bill for DC energy, not AC energy, so DC conversion losses aren't on the bill in the first place, though I assume DC chargers are also more efficient at power conversion than OBCs in general).