The solution space
Once the problem is identified, you have to come up with a solution.
The most common mistake I see is to treat the product codebase as the playground for finding the solution to a problem. Finding the right solution requires iterating an order of magnitude faster than you can deliver production software.
Plus, the uncertainty of discovering a solution makes it hard to stick to the discipline required to manage swarms of agents without producing slop.
As a consequence, the solution space must be addressed before even thinking about specs or production software. You must design a solution that is feasible, viable, valuable and usable1 outside of the context of your dark factory.
Footnotes
-
These are the four big product risks described by Marty Cagan in Inspired: value, usability, feasibility and business viability. ↩