TL;DR: Have an idea of a product. Did some research, building it as simple as glueing together a few open source SDKs with some fine-tuning, plus some devops work, should be quick. It appears to be very valuable, and sexy too.
So i understand it smells bad: i can't be the only person who thought of this, and if it doesn't exist yet, it must be a terribly bad "tarpit" idea that's certainly not worthy of working on.
How to find out why in particular is it unworkable? Any repeatable approaches to do research like that?
Consider that it may not be technically unworkable, but that there may be something that makes it less than desirable to others even if it appears valuable and sexy to you. Have you done any market research? Is there a patent or something that might be involved? Etc.