Context matters. Why you're there matters. In a business context, there's always an itching problem or a potential value creation and you can frame it in those

Context matters. Why you're there matters. In a business context, there's always an itching problem or a potential value creation and you can frame it in those terms and timebox it (drop it after X time working on it and write about the results of your tests). As much as possible, you should have a plan B that is workable even if suboptimal. On the "problem solving itself". Just break down the problem as much as possible to ensure you're slowly removing risk and learning early if there's a problem. That helps with the paralysis since tasks should feel tryable. If you can't even think about the problem then you clearly don't know enough about the domain and some separate theory and practice should help.