There is are 2 assumptions here, that needs to stated more loudly.
1: "Companies maximize profits. Humans maximize utility. Profits != utility"
2: "Profits are absolute. Utility is perceived."
Utility is an under-defined combination of components on Maslow's pyramid. Even more vexing, is the realization that these components vary from concrete (health, food, shelter) to hedonistically moving targets (social acclaim, money, material wants) to merely perceived (self-actualization). The term: "under-defined combination" is doing a lot of heavy lifting here. Because some utilities are uncomparable. 'Time spent with your children' cannot be quantified in corresponding wealth or social acclaim. Till date, I am yet to find a better compact representation (meme) of the human condition than Maslow's pyramid.
However, to a more technical audience I prefer the 'RL agent' analogy. A human is an RL agent in a weakly observed universe. The utility function it weakly defined, expected outcomes vary from true outcomes and you can only control your action space. As any RL researcher will tell you; an agent tries to maximize their utility. However, they must periodically update the utility function itself and the manner in which the world is observed. This directly applies to humans. Every once in a while, use previous experiences to re-evaluate your perception of the world. Reflect on the very foundational drivers of your goals with this new perception in mind. Find new goals. In between these iterations, a person focuses on reaching whatever their current perception of the goal is, until the next iteration.
This RL model is incredibly effective. If each iteration leads to massive changes in goals, then 'regularize' or use 'RMSE' on your utility function. If each iteration sees too little change and your utility isn't good enough; then you're stuck in a local minima. Increase your learning rate and lean more on exploration (try something new) instead of exploitation (staying your course).