Progress so far: looked into eating the lunch of an enterprise software vendor who's software doesn't actually work in a lot of cases, and after a couple evenings looking around realized there's already plenty of competition there. Competition doesn't mean don't do it, but the market was more mature than I'd realized. Given that it's fairly complicated software and even an MVP (minimum viable product) would be a ton of work for a single person, it wasn't the obvious direction I'd thought it was.
Also, the way that the thing I would've built would be better, would be "works even if your HTML markup is crap". The employees who'd use the software and might request it usually won't know enough to know that the problem is their company's markup is crap, it just looks like "other vendor is crap". The decision makers who sign off on buying usually don't know the markup is crap *and* might be insulted if it's brought up. So marketing that thing would've been an uphill battle for that reason.
The underlying technology of that vendor's product is both kind of cool, and something I think I could do better than them, though.
Then I thought of a different, simpler use case for the same tech. Basically a thingey to solve the problem where if you document a software with videos, your documentation will be out of date in months or weeks because software changes. So I think I could build the highly updatable video maker, and keep the option of expanding later to eat that bad vendor's lunch. (I have a personal beef with that company, and vengence is great motivation.)
I couldn't find anything like my idea through Google, which was *worrying*. If no one has gotten far enough in trying to solve a problem to make software about it, or even a spreadsheet or complaining blog post about it, maybe I'm misunderstanding and it's not actually a problem that matters to people.
So after work yesterday I posted on a couple different discussion boards for people who make "how to use my company's software" videos and asked if they'd want to use a thing like that, and if anything like it already existed.
No one said "I want that thing". But I do have the names of 5 companies that do the thing, do something similar, or are saying they do the thing with AI and a shitton of VC money but are rumored to not actually be doing it yet.
Next steps:
1) Talk to the makers of "how to use my company's software" videos who I actually know personally, and talk about this in more depth, possibly prepared with UI mockups that show what using the the thing would be like. Conveniently, through my spouse I know several people who do this stuff for work *and* used to do very involved tech support. I think talking to them will be really helpful for figuring out how to translate from talking about HTML to explaining in a way that'd make sense more straightforwardly.
2) Make some nice UI mockups, write about the idea in a way that makes sense to most of the people who might use it, and put together a website. Have an option to put in your email and get notified when the thing is built, and an optional text field so people can tell me stuff if there's a feature they want, if they're willing to gripe to me about what the most annoying parts of making software videos are, etc.
And after that, start coding the thingey.