I recommend starting with one of the major use cases, implementing its UI and some of its back-end, even if you have to mock some of its functionality or simpli

I recommend starting with one of the major use cases, implementing its UI and some of its back-end, even if you have to mock some of its functionality or simplify it a bit (e.g. only implementing the "happy path"). This has many benefits, not only because "management can see" the progress, but also because it's not uncommon that when we actually visualize how it will look like, our understanding of how to address the use case improves. So this reduces the requirements risk of the project, and could save you having to change or throw away back-end code at the beginning of the project. It also gives you a vertical slice on the system, so you'll have an idea of what the architecture looks like in practice (e.g. if it's too complex), which reduces technical risk.