Error committed neglecting a term

Question

I have such an expression: $$ |G|=\frac{R_2A}{R_2+(1+A)R_1}$$ (this is the gain function of an inverting op amp, but it's not important, I think..) It's noticeable that: $$|G_{\infty}|=\lim_{A\to \infty}|G|=\frac{R_2}{R_1}$$ Clearly, in real world, $A$ is big, but not infinite, so I'm wondering: what is the percent difference between $|G_{\infty}|$ (which I measure) and $|G|$ (which I should get)? I was thinking about: $$\frac{|G_{\infty}|-|G|}{|G|}\cdot 100=-\frac{R_2R_1^2+R_2^2R_1}{R_1^2R_2+R_1^2R_2A+R_1R_2^2} \cdot 100 \Rightarrow \frac{k_1}{k_1+A\cdot k_2}$$ Then, how should I proceed? I can get a percent variation at any fixed value of $A$, for every $(k_1, k_2)$ points? Is there any other extimation that exclude at least the values for $(k_1, k_2)$?