AI as augmentation.

AI tools accelerate preparation, exploration, synthesis, and documentation across every stage of the design process. They reduce friction on tasks that benefit from language generation, pattern recognition, and rapid iteration.

Important: AI-generated assumptions are not user research. Key findings and decisions must remain grounded in evidence and reviewed by people. AI augments judgment — it does not replace it.

Discovery & Research

  • Workshop planning

    Draft facilitation guides, assumption registers, and HMW prompt lists.

    Examples: Claude, ChatGPT, Gemini

  • Interview preparation

    Generate interview guides, screener criteria, and discussion prompts from research briefs.

    Examples: Claude, ChatGPT

  • Market and competitive research

    Quickly summarize competitive landscapes or market context before primary research.

    Examples: Claude, Perplexity, Gemini

  • Transcript synthesis

    Identify themes, patterns, and quotes across large volumes of interview recordings.

    Examples: Claude, Otter.ai, Dovetail

  • Insight generation

    Cluster raw notes into themed insights and generate opportunity statements (HMW).

    Examples: Claude, Dovetail, Notion AI

Design & Prototyping

  • Ideation and concept generation

    Generate alternative design directions, component patterns, and interaction models.

    Examples: Claude, ChatGPT, Figma Make

  • UX writing

    Draft button labels, empty states, error messages, tooltips, and onboarding copy.

    Examples: Claude, ChatGPT, Figma

  • Accessibility review

    Identify potential accessibility issues in designs before usability testing.

    Examples: Claude, Figma plugins

  • Edge-case exploration

    Identify scenarios, error states, and edge cases not yet covered in the design.

    Examples: Claude, ChatGPT

  • Prototype content

    Populate prototypes with realistic data, names, dates, and copy.

    Examples: Claude, ChatGPT, Figma Make

Testing & Validation

  • Test plan creation

    Draft usability test plans, task scenarios, and success criteria.

    Examples: Claude, ChatGPT

  • Findings synthesis

    Cluster session notes and identify severity-rated usability issues.

    Examples: Claude, Dovetail

  • Report writing

    Draft structured findings reports and stakeholder summaries.

    Examples: Claude, Notion AI

Development & Launch

  • Design specification

    Generate annotation content, edge-case documentation, and component notes.

    Examples: Claude, Figma

  • Analytics interpretation

    Summarize behavioral data patterns and suggest follow-on research questions.

    Examples: Claude, FullStory, Pendo

  • Post-launch evaluation

    Analyze success metric performance and generate improvement recommendations.

    Examples: Claude, Amplitude, Pendo

AI use principles for design teams

  1. Use AI to accelerate preparation and reduce blank-page friction — not to skip foundational research.
  2. Treat AI-generated insights as hypotheses to validate, not findings to report.
  3. Review AI outputs critically before using them in stakeholder communication.
  4. Document when and how AI was used in research and design artifacts.
  5. Prioritize tools with appropriate data privacy controls for your context.

AI prompts by method

A prompt for any AI assistant. It works only from material you paste in, keeps assumptions apart from findings, and leaves the decisions to you.