UI just went through the biggest shift since the invention of the touchscreen.

For decades, we’ve trained users to click buttons, fill out forms, and memorize where to find the right setting. That era is fading fast.

Today, people expect to speak, type, or gesture in natural language—and get exactly what they want.

No menus, no endless dashboards.
Just a request, and a response.

This changes everything for designers.

Because if interfaces are disappearing, what exactly is our job? How do we make products usable when the “UI” is hidden behind a thinking bubble or an AI response?

In this text, I’ll break down the three skills every designer must master to thrive in this new era:

1. More UX, less UI, as stuff happens while AI is “thinking”
2. Writing responses that are the new UI
3. Designing for time, because “AI is thinking” is the new loading screen

By the end, you’ll see why this shift is not the end of design, but the start of the most exciting chapter yet.

1. Work Closer With Developers

In the past, design was about reducing clicks.
Cleaner layouts, fewer steps, job done.
That’s not how it works anymore.

With AI, the interface is often invisible. Instead of clicking buttons, users now ask for something and expect the right output in return.

So instead of thinking “how can I make this easy to use so the user can complete their task,” we need to think “how can I take this input and turn it into the output they actually expect?”

What matters most is no longer the number of clicks saved, but whether the response is accurate, helpful, and requires fewer retries.

To design for that, you need to understand how the AI works, what happens while the user is waiting, and where the system fails. That means working closer with developers than ever before.

2. Copywriting Is the New UI

When most of the interface disappears, what’s left? The response. That’s what people interact with.

Designers need to get much better at writing, because responses are the UI now. Instead of thinking about clicks, menus, or layouts, the focus shifts to whether the output is clear, useful, and easy to act on.

Often, you only have text. So the words need to do the heavy lifting. A good response might give context, point to a source, or even suggest the next action: “Here’s the data. Want me to add a chart?” Small touches like that turn a raw output into something actually usable.

I’ve noticed this especially with ChatGPT-5. It anticipates the next step and asks a simple follow-up question to keep you moving. That works well in some cases, but if you’re just plugging an AI chat into your product, you might not want to overcomplicate it. Start by focusing on the quality of the response itself and making sure the output is actually valuable.

3. Designing for Thinking Time

When AI generates an output, users are stuck waiting. That waiting is no longer a “loading state” — it’s a “thinking state.” And how you design for that moment makes or breaks the experience.

People need to know the system is actually doing something. If they just see a blank screen or a vague spinner, trust erodes fast. ChatGPT and Claude have already improved this by telling you why it’s taking longer, or what step it’s on. That small detail changes whether a user waits patiently or bounces.

​

Designers already know this trick from skeleton loaders — gray boxes that made loading feel faster even though the time was the same. The same idea works here: narrate the thinking. Show progress like, “Analyzing your prompt… pulling sources… preparing results.”

Even better, start streaming the output as soon as possible. LLMs generate token by token anyway, so letting users see the first part of the answer reduces perceived wait time and reassures them that progress is happening.

In short: don’t hide the thinking. Design it.

Closing Thoughts

UI will always be part of software, but the way we design it is changing. Instead of focusing only on pixels and buttons, we now design for inputs, outputs, and everything that happens in between.

The skills that matter most:

• Working closer with developers to understand how the system actually works.
• Getting sharper at copywriting because responses are what people interact with.
• Making the “thinking time” feel clear and intentional so users stay engaged.

Design is still incredibly valuable. If AI is the brain behind the product, designers are the ones who make that intelligence feel usable and human.

In the end, we just need to adapt.