Polymorphicdesign
AI DesignThinking 2026
Polymorphic Design is a UX framework for creating adaptive, AI-native user experiences.
Polymorphic Design is a UX framework for creating adaptive, AI-native user experiences.
The idea
You're not average. But almost every product you use was built with one layout for everyone.
A fit for no one. It only works because you've been adjusting to it ever since.
What if we stopped?
More and more, people expect technology to reflect who they are and how they see the world.
It's time for the interface to do the adapting, and for us to just be ourselves.
UI that thinks like you, designed for you, trained by you. Tomorrow is Polymorphic Design.
Retire the average interface.
The discipline
Polymorphic Design is a way of building products that adapt to the person using them, while they use them.
Designs the standard screen and writes the rules: what can change, and what never does.
Learns how each person uses it, from what they do.
No questionnaires.
Gets a screen that fits how they think, and it keeps getting better the more they use it.
Pick an industry, then pick a person. Watch what they do, and how the same screen reshapes itself around them.
Hello
Recent Activity
Rules for what each segment sees, analytics on what people click, A/B tests, and lately some form of generative UI. Each one usually works on its own, for one segment or one request at a time. Polymorphic Design connects them around each user. You already have the pieces.
The rules set what can change and what can't, the behavior data builds the profile, generative UI draws the screen, and a group kept on the standard view shows whether the new version gets more people to the end of the flow.
Set what can change
Build the profile
Draws the screen
Prove it works
Principles
Invariance
The form changes, the substance doesn't. Price, product name and checkout options can move or get more or less prominent. They never disappear.
Contextual stability
In the same context, people see the same screen. What they've already seen changes only subtly. When the situation changes, the screen can change completely.
Gradual adaptation
Change only when the signal is clear, and change a little. Reorder and reprioritize over hours or days.
Constrained generation
A design system defines what things mean, beyond how they look. The team designs once and the AI fills it in.
Transparency
The product shows what it has learned. People can correct it or go back to the standard design.
What you sorted, dragged, opened and read on this page added up to a profile of how you take in information. It stays in your browser.
ComparerReaderExplorerSkimmer
likelihood over time
Not quite right? Pick another one.