Caution when approaching existing interfaces. After many years in enterprise UX design, if I am asked to improve an existing interface I approach it with the same respect a bomb diffuser shows for an unfamiliar bomb.

Enterprise software is full of infuriating, jaw-droppingly stupid, diabolically evil user experiences. That's one of the reasons I got into this business. When I was just a pup my reaction to these monstrosities would be "Well that's just stupid. I'll have that interaction delightful by tomorrow afternoon."

My first attempt is a thing of beauty, but requires a UI element that cannot be built using the company's current tool framework. My ingenious workaround is, indeed, ingenious, but would violate section 508 standards and expose our company to (completely justified) litigation from users with disabilities. I can still make a vast improvement by clarifying the labeling and wording; sadly my plain and simple English solution doesn't work at all in Korean.

Undaunted, I propose an exciting new visual interface. My fellow designers, who all have maps of the London subway system hanging over their beds, swoon when they see it. Actual users, however, are less enthusiastic. After watching hours of user testing I have to admit that "confused" is probably a better word. Well, confused and annoyed. One guy actually gets angry. Another woman begins to cry.

Back to the drawing board. I seek advice from PMs, developers, other designers. Gradually a workable solution begins to take shape. User testing goes much better: no one is singing but every subject completes the task without prompting. Success at last!

And then, as I prepare my presentation deck, I bring up the original UI for a triumphant before-and-after slide and discover, to my horror, that the two are almost identical. An old timer hears me sobbing at my desk and stops by. "Ah yes, we spent 18 months on that one back in 98."

NOTE: the moral of this story is NOT that you shouldn't try to fix stupid interfaces. There are many stupid interfaces that can and should be fixed. And even the ones that can't be fixed today should still be challenged; sometimes you have to lose a few battles before you can win the war. But sometimes things are stupid for a reason. Approach with caution.

In short, the main thing that separates pros from amateurs is humility.