Precision measures the proportion of correctly predicted positive instances. Accuracy assesses the overall correctness of predictions.
As a heuristic, or rule of thumb, accuracy can tell us immediately whether a model is being trained correctly and how it may perform generally.