This is a good question and there are relatively recent reviews. The progress of reactions is complex and not selective for free orthophosphate. Aquatic chemistry is most advanced in understanding the complexities.
By the way, antimony is a participant, not a catalyst.
I'm interested in the above in relation to the application of the method to soil extracts. In particular, how the lack of selectivity may relate to the poor correlation of extracted MRP values with plant growth and fertiliser requirements.