Talk to prospective customers, see what they would pay for which capabilities before you write any code. The risk is not whether you can write the application, but whether you write an application folks want to use and pay for.
I'm not so sure about this one. You can't always talk to prospective customers because they may or may not see the application of the technology until it hits them in the face. Sometimes talking to prospective customers is the right thing to do and other times you're better off building the prototype/demo and seeing what impact it has.
I don't know - maybe the OP should ask prospective customers... I don't know enough about the idea to guesstimate whether this is a good strategy or not.
