When not to use machine learning?

When you are solving a problem, in what circumstances will you apply machine learning? Is it true that in every circumstance, machine learning will always outperform rules and heuristic approaches? Before I graduated from university, I am fascinated by how machine learning models could potentially solve any problems in the world. Whenever I am facing any kind of problem, I will first think of how this could be solved by using a machine learning approach. For instance, when I was searching for a property to rent in Singapore, I would like to know the price range of one particular area.