Prototyping at the Speed of Conversation
A prototype is a question, not a deliverable. How to choose fidelity, scope, and tooling so a prototype answers the question and stops.
A prototype is a question
The most common prototyping mistake is overbuilding. A team spends two weeks building a high-fidelity prototype to answer a question that a sketch could have answered in twenty minutes. The prototype becomes a deliverable, with all the polish and none of the speed.
Every prototype should start with a written question. "Can users find the export button in the new layout?" "Does the onboarding flow make sense without instructions?" "Is the drag interaction intuitive on mobile?" The question determines the fidelity. If the question is about layout, a wireframe is enough. If it is about interaction, a clickable prototype is needed. If it is about visual hierarchy, high fidelity matters.
Choosing the right fidelity
Low fidelity — sketches, wireframes, paper — is for structural questions. Does the flow make sense? Are the steps in the right order? Is the information architecture clear? These questions do not need pixels. They need structure. Low fidelity is fast, cheap, and disposable. You should go through dozens of low-fidelity prototypes for every one that makes it to high fidelity.
Medium fidelity — clickable wireframes, gray-box prototypes — is for interaction questions. Can the user complete the task? Where do they get stuck? Is the feedback clear? These questions need interactivity but not visual polish. Medium fidelity is the workhorse of prototyping. Most questions live here.
High fidelity is for questions about the finished product. Does the visual design communicate the right tone? Does the animation feel right? Is the color hierarchy clear? High fidelity is expensive and slow. Use it only when the lower fidelities have already answered their questions.
