
Interrupting AI Is Easy. Recovering Your Thoughts Isn't.
A tiny UX detail in Google Gemini that every AI product should steal.
When people compare AI products, they usually compare the models.
I pay more attention to the UX.
Recently, I noticed a small interaction in Google Gemini that completely changed how I think about designing AI products.
It's what happens when you interrupt a response.
Imagine you've spent a few minutes writing a detailed prompt.
You hit Enter.
Gemini starts generating.
Halfway through, you remember something important.
So you press Stop.
Instead of clearing the input, Gemini puts your original prompt back into the text box.
You can immediately edit it and send it again.
It sounds like a tiny feature.
But it solves a surprisingly frustrating problem.
When we're writing a prompt, we're not just typing.
We're thinking.
The input box becomes a place to organize our thoughts.
If that text disappears, it's not just annoying—you've lost your train of thought.
Gemini avoids that completely.
This made me realize something.
As designers, we spend a lot of time making actions easier.
Fewer clicks.
Cleaner layouts.
Better navigation.
But we don't spend enough time thinking about what happens when users change their minds.
Gemini doesn't make stopping easier.
Stopping is already easy.
It makes recovering easy.
That's the part that impressed me.
This idea isn't limited to AI.
It's the same reason products have:
These features exist because people change their minds.
AI products should work the same way.
The best UX isn't always a big feature.
Sometimes it's a tiny interaction that quietly removes frustration.
Google Gemini didn't impress me because it restored my prompt.
It impressed me because it respected my thinking.
And sometimes, that's all great UX needs to do.