So we opted for a better solution: we hired a UX architect (or whatever the title is) from http://dlabs.si/ - He performed live user testing and analysis with us and delivered high fidelity wireframes, which resulted in understandable application flow, without the need for a walkthrough (or me spending time learning Kera). Granted, it wasn't free, but it was well worth the money, and it was one-off payment, as opposed to Kera's subscription model.
And as I discovered with the UX guy, it's not that difficult to perform user testing on your own, just sit down with clients and see them complete a certain task while thinking out loud along the way. It works wonders.
That being said, I'm by no means trying to discredit what you're trying to do - there certainly are use cases for products like Kera, otherwise you guys wouldn't work on it.
One thing is for sure though - the problem of "user onboarding" and conversion exists, and people are actively trying to solve it.
Hopefully the folks from iridize, tutorialize.me, walkme, taurus, etc etc will be able to take up the reigns and build a scalable business around it.