Introduction Patients diagnosed with Mild Cognitive Impairment (MCI) often go on to develop dementia, however many do not. Although cognitive tests are widely used in the clinic, there is limited research on their potential to help predict which patients may progress to Alzheimer’s disease (AD) from those that do not. Methods MCI patients (n = 325) from the longitudinal Alzheimer’s Disease Neuroimaging Initiative (ADNI-2) dataset were tracked across a 5 year period. Upon initial diagnosis, all patients underwent a series of cognitive tests including the Mini Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA) and Alzheimer’s Disease Assessment Scale-Cognitive (ADAS-Cog 13). Twenty-five percent (n = 83) of those initially diagnosed with MCI subsequently developed AD within 5 years. Results We showed that those individuals that progressed to AD had significantly lower scores upon baseline testing on the MMSE and MoCA, and higher scores on the ADAS-13, compared to those that did not convert. However, not all tests were equivalent. We showed that the ADAS-13 offers the best predictability of conversion (Adjusted Odds ratio (AOR) = 3.91). This predictability was higher than that offered by the two primary biomarker Amyloid-beta (Aβ, AOR = 1.99) and phospho-tau (Ptau, AOR = 1.72). Further analysis on the ADAS-13 showed that MCI patients that subsequently converted to AD performed particularly poorly on delayed-recall (AOR = 1.93), word recognition (AOR = 1.66), word finding difficulty (AOR = 1.55) and orientation (1.38) test items. Conclusions Cognitive testing using the ADAS-13 may offer a simpler, less invasive, more clinically relevant and a more effective method of determining those that are in danger of converting from MCI to AD.