We Won’t Have to Learn Interfaces Anymore

When we start using a new piece of software, we don’t just learn what the product does. We also learn how the product thinks. Over time, we figure out where certain actions live, where settings are hidden, how filters work, and which screens we need to move through to get something done.

UX design has spent decades trying to reduce this distance. We build clearer navigation, better onboarding, search systems, tooltips, and simpler screens. All of these make products easier to learn, but they share an assumption that has remained largely unchanged: eventually, the user still has to learn the interface.

What makes AI interesting from a UX perspective is not simply that it can make existing interfaces a little easier to use. It may change that underlying assumption altogether. Most of the time, users don’t know which button they want to press, nor should they have to. They know the outcome they want.

Requests like “Did the error rate increase after the last deployment?”, “Why did this customer’s bill go up?”, or “Close the accounts that haven’t been used in the last three months” describe intent, not an interface. If software can understand that intent and operate the systems behind the interface, users no longer have to translate their goals into a sequence of UI actions themselves.