Mild cognitive impairment (MCI) is considered a transitional stage between normal ageing and Alzheimer's disease (AD), but not all MCI cases progress to AD and there has been limited focus on how to identify who will progress. Given claims for a characteristic kind of memory impairment in AD involving deficits in encoding and consolidation of information, we propose that ‘memory profiling’ of individuals with MCI may help identify which individuals will progress. We initially set out to establish whether the same characteristic memory profile was present prior to the onset of AD (preAD). Very few studies provided data that allowed us to examine this, but results tentatively supported an encoding/consolidation profile in preAD. A single study tested the clinically important contrast of preAD versus non-preAD MCI cases and found no difference under any condition or in memory profiles, but interpretation of the findings is limited by short duration of follow-up, ceiling effects, and task limitations in assessing more complex and qualitative aspects of memory. Although existing data lead to equivocal conclusions, we believe that memory profiling is an endeavour worth pursuing, particularly given the increasing number of people with MCI presenting for clinical assessment. We propose that tests designed specifically to measure memory processes should be sensitive to preAD and are required in prospective longitudinal designs to identify these clinically crucial MCI cases.