If two people conduct a mutually-agreeable exchange, then both of them think the world has improved according to their own preferences. The employee has more money, and the employer has less money, but the work they wanted done is done. Often enough, everyone else in the world is fairly indifferent, and so pairwise mutually-agreeable exchanges often improve the world for everyone.

This is one of the ways that microeconomics is awesome. Pareto-optimality isn't everything; we do have preferences regarding things like fairness, and there are positional goods and externalities, but fairly often, working toward the Pareto frontier is a good thing to do.