Ask HN: How do you know whether or not customers will pay for your product?

Are they already paying for something which solves the same problem? If not, it probably isn't a problem.

"Something" doesn't have to be software. It could be "a manual process which consumes staff time", "a real product which could be made unnecessary by software", etc.

(Many car manufacturers solve computer problems with warehouses full of very expensive bits of steel shaped in particular ways. This is so demonstrably suboptimal that the company which licked those computer problems -- using a variety of systems which owed more to algorithms and paper cards as a transport mechanism -- basically took over the world.)

A way I have used is to create a descriptive landing page with email capture, create google ads and/or post in relevant forums to direct to page and see how many signups you get. That should give you an idea of the rough conversion rate.