At the risk of sounding like I'm patronising:

Useful Product, Well Marketed.
Don't try the other way around. You might win the sprint but you'll lose the marathon.

First thing you need is a few more heads than yours to bounce ideas around, you need to recruit a few people who believe in your idea. It helps to have something written or a simple website to pitch it to them.

Spent a lot of sweat and thoughts into the usefulness. Create usage scenarios and break them apart with your team. You'll see a lot of new ideas come out.

Then create user workflows and again, break them apart with your team. This process will give you a clear vision on your product but also allow you to create a proper development backlog.
At this point you'll be able to create a roadmap. On the short term focus on what's possible; but make sure to include a bold larger vision.

Then put all of it into a whitepaper, build a basic website and come back here for feedback. 

All kinds of new p2p dapps. Any p2p services/communities where there's trust required yet no hard coded ways to enforce it is ripe for disruption.

For example I can write from experience with p2p delivery / logistics networks: They are really crappy right now. Because they rely on old way of keeping trust by policing the community. Which can't really be done.

We're working on fixing that with project MuleChain, and we feel that the ability to add a utility token which has value in a p2p network enables the protection from abuses to be built in the system. (vs none and attempt at enforcing.) This means huge p2p logsitics networks are now possible on paper. They weren't before.

There's also probably a lot of p2p dapps categories that are yet to be born too. Things that haven't been attempted in building in p2p even pre-blockchain.