You can't even make a start.

Endless discussions, continual requests for change, but nothing ever happens.

Requirements conflict and there doesn't seem to be a way around them.
One person's bug is another person's feature.
You've built prototypes, but each one falls short of what is needed. If it satisfies one requirement, it's impossible to make it satisfy another.

Chances are you are too close to the problem. You understand your business better than anyone else can. But your requirements don't specify what your problem is, they specify solutions to your problem and the problem itself is never mentioned.

I can be an idiot. I can ask, "What is this for?", "Who is going to use it?", "How it is going to be used?", "What problem is this supposed to solve?", "What problem, that you have already solved, does this unsolve?". And most important, "How does this help you earn a crust?" And when I get the answers I can ask, "But why, Mummy, why" like a fractious three-year-old until everyone screams in frustration. But the end result really exposes the problem and allows a wide varienty of solutions.

I have done this many times for medical devices, digital communications switches, a workforce management system and astronomical instruments. I have kept my hand in by doing it as a matter of course for new or revised subsystems within existing products.