Although an automated essay-scoring engine does evaluate the Analytical Writing Assessment (AWA) sections of the GMAT, it isn't the only source for your final AWA score.
According to the official GMAT Web site, the essay-scoring engine "is an electronic system that evaluates more than 50 structural and linguistic features, including organization of ideas, syntactic variety, and topical analysis." Your two GMAT essays are also assessed by a real person — usually a college or university professor who has been trained as an AWA reader.
The live and the electronic assessors both calculate a score between 0 and 6 for each of the two essays. Those two scores are then averaged together to produce an overall writing score from 0 to 6 in half-point intervals.
If the two scores differ by more than one point, another expert reader is called in to evaluate the essays and resolve the discrepancy. Generally, this means that if the human and computer judges don't agree, another human is brought in to complete the grading process — good news for those who don't think computer programs are linguistically up to snuff yet.