One test I like: ask people to pay or pre-order before building more. A demo that impresses is not the same as something people pay for.

Agree on asking for money early. I'd add: watch whether people come back and use it on their real work in week two, not just in the demo. And ask "what do you do today instead, and what does that cost you?" If the answer is "nothing much", it's probably a nice-to-have.

Week-two use on real work seems like a much stronger signal than demo excitement. I also like asking what people use now and what that costs them; it makes the pain concrete.

Exactly, and when the current cost is concrete it also helps with pricing. If someone spends three hours a week on it today, you can anchor the price against that time instead of guessing.

Concrete time saved gives a good ceiling, but I’d still test willingness with a small paid pilot rather than assuming all saved time becomes budget. Asking what they’d pay to keep it running after the trial could make that signal clearer.