i can't remember where i found the post but i remember an artist talking about the usefulness of photo-bashing where they would take photos or assets of a variety of elements from spires to buildings and other environmental elements and frankensteining them together to get the rough concept of what they wanted and then building them out via rough sketch over the top and going from there
if one of your concerns is consistency then one thing you could try is creating a very rough (and i mean minimal detail) map of the world or setting of your comic that way you have a vague point of reference for where things lie like major landmarks in the story or whatever other points of interest your characters will engage with. it's an approach i've tried and it's actually quite useful in helping to determine the scope or scale of your world when fleshing things out. personally my preferred point of reference for this too would be video game maps specifically with the kind of setting your looking to mimic; like the ones that i remember looking to were from games like unison league, alchemia story and zelda: breath of the wild since they had the span of terrains that i was looking to recreate for my own work and best of all covered not only all the oldish architecture vibes but also those landscapes like big forests or tall cliffs and waterfalls, open plains and so on. even sky: children of the light has been a point of reference in some regards for open spaces the point is just kind of having an idea of what you want to accomplish and then between actual places and even fictional ones taking what you need to actualize those ideas
you can also absolutely put to use 3d references but as mentioned the more detail or the larger the scope the more likely you'll have to pay and not some small amounts but in most cases it ends up being well worth the investment as you've got instant access to a full model rather than having to hazard guesses on where things go and risking having inconsistencies tho it'll depend on both how much you and your readers are paying attention to such details
just remembered another little secret/shortcut i learned from tiktok which is just draw shapes and blur. depending on the distance from the character(s) your drawing you can literally take some shapes make them look vaguely important or detailed enough to resemble something and then blur it just enough so that it's no indistinguishable but recognizable enough to resemble something
i kind of screwed around with that method myself when editing this meme thing, specifically in the first frame
