IMHO basically we know very little. One thing we do know is that they are not only critical for breaking down old biomass but are 100% non-negotiable partners for many specific angiosperms. Once you lose the soil biome, those species just don't grow. Therefore when transplanting something it can be handy to bundle a sample of nearby earth at an appropriate range of depths. This can sometimes enhance germination and growth potential for difficult species by including a fungal colony... especially in difficult biomes like denuded or pool soils where a critical fungal symbiote is generally more likely.

This technique was taught to me by propagating native plant scientists here in Australia, which has a great number of unique plant species owing to its historic isolation and consequent evolutionary independence.