There are universal detectors (e.g. CAD) with which you can compare signals of different substances. But MS detectors are not universal — meaning that high intensity of the signal doesn't necessarily mean high concentration. And vice versa.
Whether the signal is high or not will depend on how well the compound is ionized. So in case of you A & B compounds:
- A may be well-ionized and thus the signal will be high even though the concentration is low;
- B is poorly ionized and you may not even see a peak even though the concentration may be high.
To identify how much of analyte you have you first need to run A with a known concentration so that you determine the relationship between area and concentration. Then if you use the same LCMS conditions you can run A with your to-be-determined concentration and by ratios of peak areas you can determine the unknown concentration.
Note, that this calculation depends on the linearity of the signal. Meaning: if the area is smaller by $x$ compared to the signal of known concentration, then the concentration will be also smaller by $x$. This isn't often the case and you may need to run multiple known concentrations to find out if the signal is actually linear. This allows you to build calibration curve which results in an $ax + b$ equation that you can further use to determine the unknown concentration.
To sum up: you can compare the signal of the same compound in different runs (given LCMS conditions and mixture matrix were the same). But not the signal of different compounds.