Most tickets I pick up describe a solution. "Add a dropdown for X." "Create an endpoint that returns Y." "Move this logic to a service." Fine. But I've learned to ask: what's the actual problem?

Not to be difficult. Because half the time, the solution in the ticket doesn't solve the real problem. Someone downstream reported a symptom, someone upstream guessed at a fix, and the ticket captures the guess. If I build the guess, I've shipped code that doesn't help.

So I dig. I talk to the person who filed it, or the person who asked them to file it. I look at the domain. Usually the real problem is smaller and simpler than the ticket suggests. Sometimes it's bigger and harder, and then it's especially good we caught it before writing code.