I don't know if this question has a straight forward answer. I have found it surprising that some professors teach organic chemistry and especially emphasize "retrosynthetic analysis" with a minimum amount of mechanisms. In speaking to some of those students they use a table to identify functional group transformations and to match starting materials and products. For the original poster, I can understand confusion if you happen to be in a similar class.
Because this pedagogy emphasizes strategy over mechanism, it shouldn't be surprising that students might have difficulty in predicting the product of a reaction to which they are trying to use the reverse process. It is like trying to solve long division problems while being unable to multiply.
I had always been quite successful in solving retrosynthesis problems despite my lack of exposure to an educational emphasis upon them. My education was heavily mechanistic. I found this mechanistic approach has served me very very well in a long and productive industrial career. Consequently, in teaching, I tried to avoid asking students to use or suggest a retrosynthesis of any reaction they could not write the mechanism for. In my usage, that means that unless you can predict the products of a reaction, then I did not expect you to predict the starting materials for that same reaction. Although that is not exactly the question being asked, I do think what I have described does (or should) apply to the vast majority of those practicing organic synthesis today.
If you wished to be as dogmatic as I, you can take any bond and analyze it as the product of radical couplings, an electrocyclic reaction, or the two polar opposite ionic couplings. You can then plug and chug your way through all of the reagents that meet those requirements. Obviously, you need to know the mechanisms in order to know which reagents can be radicals, anions, cations, and their equivalents or precursors.