The probability that material A is converted to B by a neutron depends on the energy of the neutron. The probability that B is converted to C also depends on the energy - but in a different way. There can be an energy range where the transition of A to B is likely but B to C is unlikely. That's where you want your neutrons.
But realistically, removing the material often enough is easier. Let's say you convert 1% of A to B in a week and 1% of B to C. Start with 100% A. After a week you'll have ~99% A, ~0.995% B and ~0.005% C - you lost less than 1% of your B to further conversion as B was in for less than a week on average. After two weeks you'll have ~98% A, ~1.98% B and ~0.02% C. Losses now reached 1%. After 10 weeks you'll have ~90% A, ~9.5% B and 0.5% C. Loss is now 5%. That's probably still fine. Separate A, B and C, put A back into the machine, use C elsewhere if possible or store it as waste.