Ah, in your original post you said amine and I assumed aliphatic. If you are trying to form aniline type amides, I don't think the Al - complex will work that well if at all.
Have you tried to clip the ester to form the carboxylate, remove the solvent, dissolve/suspend in something like DMF and do the amide coupling directly (HATU ect)? It avoids the whole extracting a zwitterion problem all together. I did something almost exactly the same a few weeks ago, with the exception that I used NMP as solvent and ended up using T3P to do the coupling.
Trying to go directly from the ester to the amide usually requires pushing, even more so with something as non-nucleophillic as an aniline. Even forming amides with relatively nucleophilic aliphatic amines from a methyl ester requires heating, which in the presence of base will almost certainly cause problems for your stereocenter.