The main radical in a character usually dictates how it’s read. General language familiarity tells you which of the readings to use. That’s accurate most of the time, and when it isn’t there’s furigana on the word.

For example, 青 is read as “sei”, and characters that use it as a radical are either read as “sei” or “jou”, such as in 情熱(jounetsu) or 清潔(seiketsu). So when you run into a rare character in a word that uses this same radical, you can assume that it uses a standard reading. For example, the word for fairy, 精霊, isn’t one you run into very often, but when you do you can assume that it’s read as “seirei” based on the radicals, and you’d be correct.

I’m explaining this in length here but with native level proficiency this process happens instantly, as you’re reading.

Japanese should not drop kanji. The only people that think that are foreigners that failed at learning the language. This is not a shared sentiment among japanese speakers.