I once tried the following approach and it saved me a ton of development time. In reality, it validated that there was not a good pattern for a problem to be solved in the space I was looking at.

Figure out who your target audience is first. You might have to talk to a lot of people. Get good at listening and asking open ended questions ( see book Mom Test )

Once you have a rough idea who your audience is, try to find about 30 people in that audience to talk to face to face. You can use a script like in the book Running Lean. You want to keep good notes on each conversation. Build a table/matrix of all things you uncover across these 30 or so conversations. If you can see a clear pattern in the problems a majority of the people have, that could be the idea you build a product for.