The purpose of this is to measure voltage and current from a 9V battery and the percentage of power remaining in the battery.

Well, how did you end up using 10k resistors? That won't load the battery very much. A weak battery will still show almost full voltage with that.

Also, to measure the remaining power, you must employ some mathematical battery model based on the voltage and current. The voltage is a function of current. So varying the voltage and current will not get you a single reading, it will give you a function (a curve). You can't do that with a simple voltage divider. You have to have a current sink, vary the current, and measure the resultant voltage.

That would give you a very good picture of the battery's health (you would still need a profile for the battery, because types differ).

On the other hand, a simple battery tester approximates those readings by choosing a point along the curve, using a fixed current sink (a resistor). So your talk about measuring the current is a huge "red herring" (irrelevant fact considered to be of paramount importance).

You really need to process this concept before you go on. If you really expect to measure battery capacity with 12 bit accuracy, the simple resistive divider will be a complete failure. In fact, even the first method can't be that accurate.

Also, there is some confusion in this thread about battery life. Are you considering making periodic measurements, or one measurement, or continuous measurement? Because the measuring device is measuring its own battery, then current consumption will be an issue. If not, then not really.

Perhaps I should have asked you for details of the implementation, rather than the purpose.