Overview
How might we let people iterate on an answer without retyping the whole prompt or losing a good draft?
When users often tune length, tone, or depth on the same output
- Chat and writing assistants where users commonly ask for shorter, longer, or different tone on the same answer.
- Content generation where one-click transforms beat crafting follow-up prompts.
- Products like Gemini or Notion AI that expose modify-response actions on answers.
- Flows paired with regenerate, where refinement applies a directed edit to the current draft.
When refinement chrome adds clutter without real iteration
- One-shot factual lookups where refinement chrome is unused clutter.
- Code patches where structured diff review is safer than vague “make it better.”
- When the product already supports natural-language edit-selection (magic edit) on the same surface.
States
Design the refinement loop on the answer itself, not only a hidden menu of generic improve buttons.
Answered
The first response is visible with contextual refine actions attached below or beside it.
Selecting
The user picks a directed transform: shorter, expand, formal tone, or try again.
Refining
The answer slot shows progress. Actions disable until the transform completes.
Updated
Refined text replaces the prior answer in the same surface. The original question stays put.
Key UX elements
The parts that must be present for in-place refinement to feel fast and safe.
Expose 3-4 named transforms on the answer.
Use specific labels like More concise or Add details. Avoid opaque Improve with no preview of intent.
Keep the full response in view.
Refinement updates the same answer block. Do not hide the prior text behind a loading shell alone.
Make each action describe the change.
Shorter, formal, expand, and retry should read as distinct transforms, not duplicate regenerate.
Show updating while the transform runs.
A brief in-place status prevents double taps and signals that the current draft is being rewritten.
Preserve the original ask.
Refinement applies to the visible output, not a new prompt thread. The user question stays anchored above.
Anti-patterns to avoid
Refinement actions that replace history with no undo or version carousel.
Opaque “Improve” with no preview of what will change.
Refining silently changes cited claims without updating sources.
Too many micro-actions that bury Copy and Regenerate.
How products use it
| Product | Implementation |
|---|---|
| Gemini | Modify response actions for length, tone, and structure on answers. |
| ChatGPT | Regenerate and edit flows plus conversational refinement turns. |
| Claude | Rewrite and style adjustments via follow-ups and artifacts. |
| Notion AI | Inline improve/shorten/change-tone actions on generated text. |
Real-world examples
How shipped products implement response refinement, from our teardown guides.
ChatGPTChatGPT output UX: in-thread replies & refinement
Output, artifacts & refinement
ClaudeClaude output UX: replies, retry & feedback
Output, artifacts & refinement
GeminiGemini output UX: guides, tables & Workspace handoffs
Output, artifacts & refinement
ManusManus output UX: agent runs, files & deliverables
Output, artifacts & refinement
PerplexityPerplexity citations UX: source chips & trust design
Trust, privacy & settings
PerplexityPerplexity output UX: tabs, follow-ups & export
Output, artifacts & refinement
PerplexityPerplexity personalization UX: vertical profiles, response style & memory
Personalization & memory
Implementation
Copy this prompt to generate a production-ready implementation in Cursor, Claude Code, Lovable, or any AI coding agent.
Generate a production-ready implementation of the "Response Refinement" AI interface design pattern.
Pattern Definition:Frequently asked questions
What is response refinement in AI UX?
Response refinement gives users actions to change an existing answer (shorter, longer, tone shifts, continue) without starting from a blank prompt, keeping iteration attached to the output.
How is refinement different from regenerate?
Regenerate samples another full answer from similar intent. Refinement applies a directed transform to the current answer (or selection). Strong products offer both.
Should refinements create versions?
Yes when users compare variants. A simple version carousel or undo prevents loss of a good draft after an aggressive shorten or rewrite.