I prefer the isolation from soil concerning soil borne but some antagonists isolats de from plant confirmed their inhibiteur effect against some pathogens.
I am doing research similar with the topic you questioned.
i prefer to isolate fungi from the plant (especially roots part). root-colonizer fungi is best to compete with soil-borne pathogen. For air-borne pathogen also. because this fungi who live inside plant can activate plant ISR.
for that, you need to examine rheir ability to colonize plant root by inoculation test to the plant
Soil isolates of biocontrol agents preferably from rhizosphere is better than any other part of the plant. Its competitve saprophytic ability (CSA) is more.
In case if plant deseases it id preferbly to isolate pothogen from plant such as Fusarium.if you dont to Know the antagonist effect it id better from soil
Hi! I recommend to isolate them from the soil. It's a slow process because you have to test the pathogenicity against insects. In that case, you can use Galleria mellonella larvaes as a hosts. But you can't be full sure that you are goint to isolate only fungi. You can also isolate Entomopathogenic bacterias or nematodes. It's a really good option to use the natural enemies as biological control agents because of their adaptation but if there are several damages in the crop, I prefer use non native in order to reduce the pest population.