GlossaryPrompting and interaction

Few-Shot Prompting

Few-shot prompting includes a small number of example input/output pairs in the prompt so the model mimics your format, tone, or decision style.

Products embed few-shot examples invisibly (starter templates) or visibly (“See examples”) to raise first-reply quality.

What it means

You show the model one to several demonstrations of the task done correctly, then ask it to handle a new input the same way.

Why designers should care

Few-shot is a design lever for consistency (release notes, support macros, UX microcopy, critique rubrics) without retraining the model.

Example

A UX writing assistant ships with three approved before/after microcopy pairs in the system layer. Users pick “Error message” and the model matches your voice on the first draft.

Common mistakes

  • Examples that contradict each other or drift from brand guidelines.
  • Too many long examples that eat the context window before user content fits.
  • Examples visible to users but not editable, causing mistrust when output diverges.

Weekly AI UX notes

Patterns, prompts, and glossary updates for designers building AI products on Substack. No spam.

Subscribe on Substack