How Designers Can Harness ChatGPT Throughout the Entire Design Journey
This article chronicles a designer's hands‑on experiment using ChatGPT as a collaborative partner across a design‑driven project, detailing how AI‑generated questions, market analysis, strategy formulation, and execution support can boost efficiency while highlighting the model's current limitations in visual input and nuanced decision‑making.
Since ChatGPT’s launch, new updates keep emerging. As a designer without a technical background, I was curious about walking the entire design journey side‑by‑side with ChatGPT and documenting the experience.
Project Background
ChatGPT is a language model that currently only accepts text input and output, so we used it within a design‑driven project to test its assistance at various stages.
Recommended Inquiries
Communication outline (product goals, positioning, stage, planning, target users, scenarios, etc.).
Key Help
More comprehensive interview planning, reducing repeated communication.
Notes
Often need to add expected timelines and cross‑department collaboration details.
We created a spreadsheet for six designers to write questions, merged them, and also asked ChatGPT for supplemental questions.
ChatGPT provided ten questions, lacking product‑specific input, but it added valuable marketing and promotion ideas.
Analysis of Questions
Recommended Inquiries
User analysis (target traits, acquisition channels, research methods, questionnaire), data analysis (metrics, methods, evaluation dimensions), competitor analysis (names, methods, pros/cons).
Key Help
Direct answers for user and competitor issues in a saturated market; data analysis suggestions for more comprehensive metrics.
Notes
Clearly state the industry and specific goals; otherwise answers become too generic.
We discovered that adding restriction words (e.g., “in China”) improves relevance, and including domain‑specific terms (e.g., “live‑stream data metrics”) yields tailored responses.
For vague questions like “what product metric is considered qualified,” ChatGPT explained that high numbers don’t necessarily mean success and suggested ways to evaluate metrics, providing a conversational feel.
When faced with questions beyond its knowledge, ChatGPT avoids a direct “I don’t know” and instead offers plausible‑looking but sometimes inaccurate answers, such as mischaracterizing a recruitment live‑stream platform.
Solution Strategy & Design Execution
Recommended Inquiries
Solution strategy (design model, approach) and design execution (resources, copy, components).
Key Help
Excellent copywriting; other tasks can be handled like a search engine.
Notes
ChatGPT‑4’s image input is future work; for now graphic output can be assisted by image‑focused AI tools.
Because ChatGPT cannot quickly grasp background, constraints, or produce graphics, its overall assistance is moderate, though it still helps expand ideas.
When looking for specific design models, adding concrete examples (e.g., “AIPL model”) guides ChatGPT to suggest more relevant frameworks.
Even without visual output, ChatGPT can recommend suitable UI components for conveying information.
When Is ChatGPT Most Effective in Design?
It shines in saturated markets with mature methodologies where accurate information can boost efficiency and even provide direct answers.
However, its content relies on “search and stitching,” lacking foresight and decision‑making, making it unsuitable for interpersonal management and strategic decisions that require nuanced human judgment.
Design Thinking & Innovation
Thinking involves both rational reasoning and non‑rational psychological processes such as innovation, problem‑solving, analysis, synthesis, and evaluation.
Innovation is debated; some claim AI cannot innovate, yet examples like AlphaGo’s creative moves show AI can produce novel combinations of existing elements.
Human designers bring lived experience and insight, enabling truly novel solutions that AI struggles to replicate.
For instance, deriving metaphorical UI expressions based on tool usage remains challenging for AI.
When designers lack ideas, asking ChatGPT can be more efficient than a generic web search.
Nevertheless, designers should avoid treating themselves as robots; many tasks are still better performed by humans, and the unique value of human designers lies in contextual understanding and creative judgment.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
