A good estimate would be the B1-B2 multipliers in AISC. There is not a set percentage. If there were, it would be put in the code and we could avoid the detailed analysis.
Just want to know a common situation. For example, one would like to control KL/r around 90-150 for column.
p-delta effect was some thing we can control by change the
rigidity of structure, as I understand. Similar as KL/r, I think it may have a allowance value.
I usually look more to the total deflection (or drift) than to the amount of the p-delta increase in moment. I have done water towers where the p-delta effect with seismic has added over 50% to the design moment.
I don't know off-hand what it would usually be for buildings. I don't really consider them without p-delta to know the difference. It would be proportional to the gravity load and the drift - both of which can vary from job to job.
I'd estimate that a total amplification of about 5 to 10% is common for building type structures. But, that is just a ball park number for those of us in seismic country.
This can increase if there are very few moment frames with a lot of "leaning" columns that don't participate in resisting the lateral forces.... more common for structures where the lateral design is entirely controlled by wind.