Step 1: The task is prediction: given x, estimate y. That is the job of regression, not correlation.
Step 2: The coefficient of regression quantifies how much the dependent variable (y) changes for a unit change in the independent variable (x), so it lets you compute the actual value of one variable from the other.
Step 3: The coefficient of correlation only tells you the strength and direction of the association, not the predicted quantity. The coefficient of variation is a measure of relative spread, and the coefficient of determination (r squared) tells the proportion of variance explained, not the predicted value.
Ref: Fundamentals of Biostatistics (regression and correlation).