Generally, I'd recommend figuring out what lightning's advantages are over tradfi and look for problems that can be addressed using those advantages.

e.g. What businesses exist already where people don't get paid, or the market is incompletely addressed, because

Another framework: What product suffers severely from a lack of sybil resistance?

Another: What product would benefit from instant settlement?

You can also address internal bitcoin problems, sort of creating a shovel for people building such things.