I don't know if I can put enough into a TL/DR; to even be useful, but the general gist of things starts with something like:
Have some kind of idea of approximately what you're interested in building, or what kind of problem you are interested in solving. If you're really starting from a complete blank slate, this might be tough. It doesn't need to be super detailed, but you should have something to start from. Otherwise you're asking people really airy questions like "What problems do you have?" that don't always work well as a starting point.
Talk to people in industry who you already know (Blank calls these "Friendly First Contacts") and ask questions around their experience in the area related to your vague notion of what you want to build / what kinds of problems you want to solve. Next, and perhaps most importantly of all... ask your "friendly first contacts" to refer you to other people that they think could give you useful insights, but who you don't already know. Repeat this process (and "the process" as laid out by Blank is fairly detailed, structured, and actionable. This is not hand-wavy mumbo jumbo. But you'll have to read the book(s) for the deep details).
As part of all this, you employ an iterative process of developing a detailed "problem hypothesis" which is simply a statement along the lines of "companies (or people) like you often have this problem, and would spend money to solve it." You then go back to your contacts (while also continually working to expand that pool of contacts, as you see fit) and try to validate / invalidate your hypothesis. If you can't validate your hypothesis, you start over and develop a new hypothesis. Once you have a problem hypothesis that you feel is validated (to some extent. This isn't really a totally binary thing) you develop a "solution hypothesis" which says "given that you have this problem, a product like the following would solve the problem and you would pay for it." And you keep iterating, talking to people, presenting your hypothesis, trying to validate/invalidate it, etc.
If it all works, at some point you have good reason to believe that you know something substantial about a real problem that real people have and a real (potential) solution for that problem that they would (theoretically) pay real $$$ for.
I'm eliding a lot of details, but somewhere in all this is where you build an MVP, show it to people you've talked to, and ask questions like "If I built this product would you pay a million dollars for it?" and "If I offered this product to you for free, would you install and use it immediately?" and others to help you hone in on just how real the interest and potential are.
Somewhere after this is where you start asking for either a check and signed purchase order (up until now you haven't been trying to sell anything), or at least something like a non-binding letter of intent that you can use to demonstrate the reality of interest in your solution if you're going to go out and try to raise outside capital, etc.
There's more to it than this of course. But at a super, super high level, that's sorta kinda how you can approach things. HTH.
[1]: https://www.amazon.com/Four-Steps-Epiphany-Successful-Strate...
[2]: https://www.amazon.com/Mastering-Complex-Sale-Compete-Stakes...
[3]: https://www.amazon.com/Prime-Solution-Increase-Margins-Compl...
[4]: https://www.amazon.com/Mastering-Complex-Sale-Compete-Stakes...
[5]: https://www.amazon.com/CustomerCentric-Selling-Second-Michae...
To summarize part of the approach: you want to have high-bandwidth conversations with specific individuals who you think are part of your target market. Make sure you and they agree on the problem. Then try asking a question like "If I built a solution to X and offered it to you for free, would you deploy it?"
If the answer is "no" then it's obvious that the problem isn't a particularly high-value one, OR you don't actually have a proper shared understanding of the problem.
If the answer is "yes" then you can start trying to bound the actual value of the problem. "Well, what if I charged you one million dollars for the solution? Would you deploy it then?" The answer may well be "no", but that's OK. Now you can start a sort of a binary search process to try and find some common ground.
Beyond that, there's a series of books by a guy named Jeff Thull that I really like, which you might find valuable, at least if you're trying to sell something in a B2B setting. The first book in the series is titled Mastering the Complex Sale[2] and a big focus of Thull's approach is about how to develop a "shared understanding" of a problem, and the value of a solution. It's a "sales" book, but it's not the kind of "sales book" that's about trickery and manipulation, etc. It's really about how to work together with a prospect to gain that "shared understanding." I'm a big fan, FWIW.
[1]: https://www.amazon.com/Four-Steps-Epiphany-Successful-Strate...
[2]: https://www.amazon.com/Mastering-Complex-Sale-Compete-Stakes...
https://www.amazon.com/Four-Steps-Epiphany-Successful-Strate...
There's also no real point in running ads when all you have is an MVP. Ads are for demand generation once you have a really solid handle on the problem, the product, sales channels, etc.
My suggestion would be the following:
1. Stop writing code or doing any development work
2. Cut expenses to the bone by any means necessary so that you're "default alive" no matter what (for the foreseeable future anyway)
3. Get yourself a copy of The Four Steps to the Epiphany[1] by Steve Blank. Read it. Then read it again.
4. Go through the customer development process as outlined by Blank, and start talking to prospects.
5. Don't start writing code again until/unless you sync up to that point where the process has you starting to take an MVP out to show to prospects.
6. Pick up from there and follow the process.
Now an exception would be this: you talk about a mature market. IF there are already plenty of well established, mature, market proven solutions that do more or less the same thing as your proposed solution, then you can skip some of the validation stuff. I mean, you could guarantee that you're at least building "something people want" by just copying one or more of your competitors. BUT... if all you're doing is copying a competitor, you have to ask "why would a customer buy your solution instead of theirs?"
Ideally in that case, you have some vision or idea for a "thing" that distinguishes your solution from the others on the market. In that case, though, you still have to validate that the market players actually see your proposed solution as better than the other solutions OR you need to be "at least as good as the others" but significantly cheaper, or something.
In terms of addressing all of that, a couple of books I'd suggest reading include:
1. The Discipline of Market Leaders[2]
2. Differentiate or Die[3]
3. Zero to One[4]
4. Traction[5]
[1]: https://www.amazon.com/Four-Steps-Epiphany-Successful-Strate...
[2]: https://www.amazon.com/Discipline-Market-Leaders-Customers-D...
[3]: https://www.amazon.com/Differentiate-Die-Survival-Killer-Com...
[4]: https://www.amazon.com/Zero-One-Notes-Startups-Future/dp/080...
[5]: https://www.amazon.com/Traction-Startup-Achieve-Explosive-Cu...