The tools exist and people will name them, HeyGen and ElevenLabs are the usual pair for voice plus lip sync, but I'd push back gently on the framing because the thing making the Japanese version feel different probably isn't the tool.
What's happening is basically that your master was edited to the rhythm of the original language. Languages don't take the same amount of time to say the same thing. Japanese and Spanish both run longer than English for identical meaning, so once you drop the new VO in, your cuts stop landing where the sentences land. The words are right and the timing is off by half a beat everywhere, and that reads to a viewer as a slightly cheaper ad without them being able to say why.
The production fix is to stop making a video that needs lip sync at all. If the master is built as voiceover over product and hand shots rather than somebody talking to camera, every language version becomes a swap of audio and subtitles and nothing else. Where you do need a face, keep the on camera lines short and self contained with cutaways between them, so each language can run long or short inside its own segment without breaking the edit.
One other thing on subtitles, Japanese needs far fewer characters per line than Spanish before it stops being readable, so a single subtitle template across all three will look wrong on at least one of them. We set line length per language rather than per project.
For skincare specifically I'd be careful with AI lip sync on close ups. It holds up in a wide shot and falls apart the closer you get to the mouth, and skincare ads live in close up.