I will give a analogy first, you know what dish you want but you don't know the recipe.
Everyone here was a novice when they started, you can ask chatgpt to generate a dalle2 prompt for your situation and it can generate a reasonable first output then you can tweak from it.
you can look at other peoples work for discord on midjourney
There are also sites which show generative images and prompt used like Lexica https://lexica.art
There are paid marketplaces for prompts like https://promptbase.com where readymade prompts are available and also people who offer prompt writing services.
How does it "failed" for you ?
Here is nice guide to learn more and practice with prompts: https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y/ae3935aa-88...
All the best - it's good fun
It generated ~0.01% of what was intended. The task is take an object from one type of images and insert it instead of another object from another thematic category of images. It tried to insert some very few elements of the first object alongside that another object, but not instead of it (didn't fully replaced it). And in one 3D render it inserted a graphical symbolic representation of this first object, rather than the object itself.