if you were going to do this with ae, i'd drop your footage into a pal standard comp.
size it to fit the pal frame (layer>transform>fit comp width or fit comp height).
then set time stretch to 104.271% -- 100*25fps/23.976fps. if you have true 24fps footage, then 100*25fps/24fps = 104.167%.
you can enable frame blending (pixel motion) for the footage layer and ae will interpolate the second field data, you you can choose to render 25i without frame blending which will the appear like 25p, either option is usually aceptable... frame blending takes longer, but usually looks smoother, however, some transitions (like cuts) can create some issue (due to trying to interpolate a new frame between two frames from different scenes -- essentially you get one garbage frame)... i would probably lean towards rendering without frame blending, but it depends on the project.
note, changing the speed of the footage will cahnge the pitch of the audio slightly... many nle's will have the ability to fix the pitch and may be a better choice for this sort of thing, but in ae, that's the method i'd use.