i'd start a business where innovation happens behind the scenes instead of out front.
in other words, not software. it's highly competitive and developers are expensive. so your "95% software margins" aren't really 95%, because you have to pay smart people to keep it running. that's no different than a nice restaurant doing 40% gross margins but only paying cooks $15 /hour.
when innovation happens behind the scenes, it's tougher to copy. things like: process, systems, documentation. these are domains a good thinker can eek out profits without the competition finding out, cloning you, and resetting the hedonic treadmill.
this has been on my mind a lot after reading Charlie Munger's reflections on Berkshire's takeover of See's Candy. he notes that some companies are capital intensive, requiring ongoing investment to stay relevant. but some companies, like See's, have used the same metal pots to make the same chocolate for over 50 years.