What it is
Using large language and image models to generate, expand, remix, and critique ideas, treating the AI as a fast, unopinionated ideation partner that widens the option space.
When to use it
During Ideate, to break fixation, generate volume, or explore adjacent directions quickly.
Why (value)
Rapidly expands and diversifies ideas, offers unexpected angles, and lowers the cost of exploring many directions before human judgment converges.
How to run it
- Frame the problem and constraints in the prompt.
- Ask for many diverse ideas, then push for wilder or adjacent ones.
- Use the model to critique or combine ideas.
- Curate with human judgment, discard the generic.
- Take promising ideas into sketching or prototyping.
Inputs and outputs
- Inputs: A well-framed prompt and access to a generative model.
- Outputs: A broad, curated set of candidate ideas.
Time and effort
Low. Minutes.
Evidence and limitations
AI output is inference from training data, not evidence about your users, and can be confidently wrong. Human curation and later validation are essential. See Evidence and limitations.
Tips and pitfalls
- Treat AI ideas as raw material, not answers.
- Push past the first, generic batch.
- Never mistake fluent output for validated insight.
Variations
LLM brainstorming, image-model concepting, AI critique and remix.
Related methods
Need help choosing and applying a method?
Gumshoe helps teams connect research, collaborative workshops and service design to the decisions they need to make.
Sources and further reading
- UX Tigers, “100 UX methods.”
Explore related methods
- Brainstorming — Generate a large volume of ideas quickly by deferring judgment.
- SCAMPER — Prompt new ideas by systematically transforming an existing thing seven ways.
- AI-assisted copy drafting — Use AI to draft and vary UX copy fast, then edit with human judgment.
