Specificity Principle

AI outputs match the specificity of the input.

How to apply it

State purpose, layout, color, typography, density, and dimensions. Reference a known style.

Specificity Principle in practice: a concrete example

Vague prompt in, generic result out

Ask an AI tool for “a nice landing page” and you get a bland, average one, because you left every real decision to it. Specify the purpose, the layout, the colour direction, the typography, the density, and the exact dimensions, and reference a style it knows, and the output snaps toward what you actually pictured. The model matches the specificity of the input: the more precisely you constrain it, the less it has to guess, and the closer the first draft lands.

Common Specificity Principle mistakes

Giving a one-line prompt and blaming the model when the result is generic. It filled the gaps you left.
Piling on adjectives (“modern, clean, premium”) that sound specific but constrain nothing concrete.
Leaving out the measurable constraints (dimensions, layout, density, a named reference) that actually pin the output down.

When Specificity Principle doesn't apply

Precision helps most when you already know what you want. Early in exploration the opposite is useful: a loose, open prompt lets the model surface directions you hadn't considered, which you then tighten. Over-specifying too soon can also fight the model, forcing a worse result than a clear goal plus room to solve it would give.

Practise it on a real challenge

AI Mockup — the prompt behind a clean pricing card

You want to use an AI image tool to mock a clean SaaS pricing card. The prompt decides the result.

Specificity Principle

How to apply it