ChatGPT Prompts for Product Designers
Thirty role-specific prompts that turn ChatGPT into a tireless design partner — from first concept to clean developer handoff. Copy, paste, and swap in your own [PRODUCT], [FEATURE], and [USER].
In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 15 prompts are free instantly — no signup needed. Hand-curated and tested by the AI Academy team.
Concept Ideation
5 promptsDivergent Concept Generator
1/30<context> I am a product designer working on [PRODUCT]. I need to explore a new [FEATURE] for [USER]. We are in early divergent ideation and want breadth before we converge. </context> <task> 1. Generate 8 distinct concept directions for [FEATURE], each solving the core job differently (do not just reskin one idea). 2. For each concept give: a one-line name, the core mechanic, the primary user benefit, and the riskiest assumption. 3. Rank the 8 by a 1-5 desirability score and a 1-5 feasibility score. 4. Flag the 2 concepts that are most non-obvious but still viable, and explain why they are worth a sketch. </task>
Produces a ranked spread of eight differentiated feature concepts with risks and a shortlist of non-obvious bets.
Pro tip: After the list, reply 'expand #3 and #7 into rough screen-by-screen flows' to drill into the winners without re-prompting.
Analogous Inspiration Miner
2/30<context> I am designing [FEATURE] for [PRODUCT], used by [USER]. I want to break out of category clichés by borrowing patterns from unrelated industries. </context> <task> 1. Identify 6 products or experiences from OUTSIDE my industry that solve a structurally similar problem to [FEATURE]. 2. For each, describe the specific interaction pattern or mental model that works. 3. Translate each pattern into a concrete design idea for [PRODUCT]. 4. Note one pitfall to avoid when adapting that pattern to my context. </task>
Surfaces cross-industry interaction patterns and translates each into a concrete idea for your product.
Pro tip: Name a beloved app in your prompt ('think Duolingo streaks') to anchor ChatGPT on the emotional mechanic, not just the UI.
Problem Reframing Partner
3/30<context> My team frames the problem as: '[USER] struggles to do X with [PRODUCT].' I suspect we are solving the wrong problem and want to challenge the framing before designing. </context> <task> 1. Restate my problem in 5 alternative framings, each shifting the assumed cause or the desired outcome. 2. For each reframe, state what new solution space it opens. 3. Identify which framing is most likely to be the real root problem and why. 4. Suggest 2 quick research questions to validate the strongest reframe before we commit. </task>
Generates five alternative problem framings, opens new solution spaces, and recommends validation questions.
Pro tip: Paste a raw user-interview quote into the context so ChatGPT reframes against real language, not your internal jargon.
Constraint-Driven Brainstorm
4/30<context> I am designing [FEATURE] for [PRODUCT] under tight constraints: it must work offline, ship in one sprint, and add zero new screens. [USER] is non-technical. </context> <task> 1. Generate 6 ideas that respect ALL three constraints simultaneously. 2. For each, explain how it honors the offline, one-sprint, and no-new-screen limits. 3. Highlight any idea that turns a constraint into an advantage. 4. Call out the single idea with the best effort-to-impact ratio and justify it. </task>
Delivers six ideas that fully respect your hard constraints, flagging the best effort-to-impact option.
Pro tip: List your real constraints explicitly — ChatGPT brainstorms far better when boxed in than when asked to 'be creative.'
Opportunity Sizing Sketch
5/30<context> I have a backlog of feature ideas for [PRODUCT]. I need to prioritize which to design first for [USER]. The ideas are: [LIST IDEAS]. </context> <task> 1. Score each idea on reach, impact, confidence, and effort (RICE), using reasonable assumptions you state explicitly. 2. Produce a ranked RICE table. 3. Recommend the top 3 to design next and explain the trade-offs. 4. Flag any idea that looks high-value but is too uncertain to start without research. </task>
Turns a raw idea backlog into a transparent RICE-scored, prioritized shortlist with stated assumptions.
Pro tip: Ask ChatGPT to 'mark every assumption you invented in bold' so you can sanity-check the inputs before trusting the ranking.
Prompts get you started. Tutorials level you up.
A growing library of 300+ hands-on AI tutorials. New tutorials added every week.
Design Briefs
5 promptsOne-Page Design Brief
6/30<context> I am kicking off design for [FEATURE] on [PRODUCT], serving [USER]. Stakeholders are PM, eng lead, and marketing. I need a crisp brief to align everyone. </context> <task> 1. Draft a one-page design brief with these sections: Problem, Target user, Success metrics, Scope, Out of scope, Constraints, Open questions. 2. Keep each section to 2-4 bullet points, written for a mixed audience. 3. Phrase success metrics as measurable outcomes, not features. 4. End with a single sentence stating the design hypothesis we are testing. </task>
Outputs a tight, stakeholder-ready one-page brief ending in a testable design hypothesis.
Pro tip: Feed ChatGPT the meeting notes from kickoff and ask it to 'fill the brief only from what is stated, leave gaps as open questions.'
Success Metrics Definer
7/30<context> We are designing [FEATURE] for [PRODUCT]. Leadership wants to know how we will measure success for [USER], but we only have a vague goal: 'improve engagement.' </context> <task> 1. Translate 'improve engagement' into 4-6 specific, measurable metrics tied to [FEATURE]. 2. For each metric, give a definition, a data source, and a sensible target direction. 3. Separate leading indicators from lagging indicators. 4. Recommend the single North Star metric and one guardrail metric to prevent gaming it. </task>
Converts a vague goal into defined leading and lagging metrics with a North Star and a guardrail.
Pro tip: Ask for the guardrail metric explicitly — ChatGPT will otherwise optimize the brief toward a single number that is easy to game.
Stakeholder Alignment Map
8/30<context> I am about to brief [FEATURE] for [PRODUCT]. The stakeholders are [LIST ROLES]. They have competing priorities and I want to anticipate friction. </context> <task> 1. For each stakeholder role, infer their likely top priority and their biggest concern about [FEATURE]. 2. Identify the 2 most likely points of conflict between stakeholders. 3. Suggest framing for the brief that pre-empts each conflict. 4. Draft one alignment question to ask each stakeholder before design starts. </task>
Maps each stakeholder’s priorities and concerns, predicts conflicts, and supplies pre-empting questions.
Pro tip: Name real people and their past objections in the context; ChatGPT tailors the framing far better with specifics than with generic roles.
Scope-Cutting Advisor
9/30<context> The brief for [FEATURE] on [PRODUCT] has grown too big to ship this quarter. I need to cut scope without losing the core value for [USER]. </context> <task> 1. Separate the brief into must-have, should-have, and could-have using MoSCoW. 2. Define the smallest version that still delivers the core user value (the walking skeleton). 3. List what we defer and the explicit risk of deferring each item. 4. Propose a phased rollout: v1 now, v1.1 next, with the trigger for each phase. </task>
Splits a bloated brief into MoSCoW tiers and defines a shippable walking skeleton with a phased rollout.
Pro tip: Paste the full brief and tell ChatGPT 'be ruthless — assume we have half the time we think' to get a genuinely lean v1.
Competitive Brief Audit
10/30<context> Before designing [FEATURE] for [PRODUCT], I want to ground the brief in how competitors handle the same job for [USER]. Key competitors: [LIST COMPETITORS]. </context> <task> 1. For each competitor, summarize how they solve the job [FEATURE] addresses (use general knowledge, flag uncertainty). 2. Identify the table-stakes elements we must match. 3. Identify 2 gaps or weaknesses we could exploit as differentiation. 4. Recommend one positioning angle for the brief and the design decision it implies. </task>
Audits competitor approaches to find table stakes, exploitable gaps, and a differentiated positioning angle.
Pro tip: Treat competitor facts as hypotheses — ask ChatGPT to label anything it is unsure of so you verify before writing it into the brief.
User Flows
5 promptsEnd-to-End Flow Outliner
11/30<context> I am mapping the flow for [USER] completing the core task of [FEATURE] in [PRODUCT]. I want a complete happy-path flow before I touch Figma. </context> <task> 1. List every step from entry point to task completion as a numbered sequence. 2. For each step, note the user goal, the screen or surface, and the primary action. 3. Mark decision points where the flow branches. 4. Identify the single step most likely to cause drop-off and why. </task>
Produces a numbered end-to-end happy-path flow with branch points and the highest drop-off risk flagged.
Pro tip: Once the happy path is set, ask 'now add the unhappy paths for each decision point' to expand without losing the structure.
Edge-Case Enumerator
12/30<context> I have designed the happy path for [FEATURE] in [PRODUCT] for [USER]. Before handoff I need to cover edge cases and error states. </context> <task> 1. Enumerate edge cases across these categories: empty states, error states, permission/auth, network failure, and extreme inputs. 2. For each edge case, describe the trigger and the ideal user-facing behavior. 3. Recommend the message or microcopy the user should see. 4. Prioritize which edge cases must be designed for v1 versus deferred. </task>
Enumerates edge and error cases by category with ideal behavior, microcopy, and v1 priority.
Pro tip: Ask ChatGPT to 'write the error microcopy in our voice' and paste a sample of your product copy so the states sound on-brand.
User Journey Mapper
13/30<context> I want a journey map for [USER] adopting [FEATURE] in [PRODUCT], spanning awareness to habitual use, not just a single session. </context> <task> 1. Break the journey into stages: Awareness, First use, Activation, Habit, Advocacy. 2. For each stage describe the user goal, emotional state, and key touchpoint. 3. Identify the biggest friction or doubt at each stage. 4. Suggest one design intervention per stage to move the user forward. </task>
Builds a five-stage journey map with goals, emotions, friction, and a design intervention per stage.
Pro tip: Add your real activation metric to the context so the 'Activation' stage maps to the moment your data actually tracks.
Information Architecture Sorter
14/30<context> I am restructuring the navigation of [PRODUCT] to better support [FEATURE] for [USER]. Current sections are: [LIST SECTIONS]. </context> <task> 1. Propose a cleaner IA, grouping content by user mental model rather than internal team structure. 2. Show the proposed nav as a labeled hierarchy (max 3 levels deep). 3. Explain each major grouping decision in one sentence. 4. Flag any current label that is jargon and suggest a clearer user-facing term. </task>
Reorganizes navigation around the user’s mental model with a labeled hierarchy and clearer labels.
Pro tip: Ask for a quick 'card-sort rationale' so you can defend the new IA to stakeholders who are attached to the old structure.
Onboarding Flow Designer
15/30<context> New [USER] sign up for [PRODUCT] but drop off before experiencing the value of [FEATURE]. I need an onboarding flow that gets them to first value fast. </context> <task> 1. Define the single 'aha moment' for [FEATURE] and the shortest path to it. 2. Design a step-by-step onboarding flow with the minimum steps needed to reach that moment. 3. For each step, specify the goal, what to show, and what to defer. 4. Recommend where to use progressive disclosure versus required input. </task>
Designs a minimal onboarding flow built around the aha moment with progressive disclosure guidance.
Pro tip: Tell ChatGPT 'cut any step that is not strictly required to reach the aha moment' — it defaults to over-onboarding otherwise.
Prototyping Specs
5 promptsInteraction Spec Writer
16/30<context> I am specifying the interactions for [FEATURE] in [PRODUCT] so I can build a high-fidelity prototype. The key component is [DESCRIBE COMPONENT]. </context> <task> 1. Write a detailed interaction spec for the component: default, hover, active, focus, disabled, loading, and error states. 2. For each state, describe the visual change and the trigger. 3. Specify transition timing and easing recommendations. 4. Note any accessibility behavior (keyboard, screen reader, focus order) for each state. </task>
Generates a complete state-by-state interaction spec including transitions and accessibility behavior.
Pro tip: Ask ChatGPT to output the states as a table so you can paste it straight into your Figma prototype documentation frame.
Microinteraction Designer
17/30<context> I want to add a delightful but purposeful microinteraction to [FEATURE] in [PRODUCT] for [USER]. The trigger is [DESCRIBE ACTION]. </context> <task> 1. Propose 3 microinteraction concepts for this trigger, each with a clear functional purpose (not decoration). 2. For each, describe the feedback, the animation, and the duration. 3. Explain the user benefit and the risk of overdoing it. 4. Recommend the one most appropriate for our context and why. </task>
Proposes three purposeful microinteractions with feedback, timing, and a context-fit recommendation.
Pro tip: Specify your product’s personality ('calm and professional' vs 'playful') so the animation suggestions match your brand’s motion language.
Component Variant Spec
18/30<context> I am building a [COMPONENT] for the [PRODUCT] design system, used across [FEATURE]. I need to define its full variant matrix before building in Figma. </context> <task> 1. Define the relevant properties for this component (e.g. size, variant, state, icon). 2. Produce the full variant matrix as a table of property combinations. 3. Flag combinations that are invalid or should be blocked. 4. Recommend sensible defaults for each property. </task>
Defines a component’s property set and full variant matrix, flagging invalid combinations and defaults.
Pro tip: Ask ChatGPT to name the variants using your design-system convention so they map 1:1 to Figma component properties.
Prototype Test Script
19/30<context> I have a clickable prototype of [FEATURE] in [PRODUCT] and need to run unmoderated usability tests with 5 [USER] participants. </context> <task> 1. Write a test plan with a clear objective and 3-4 realistic tasks for [FEATURE]. 2. Phrase each task as a goal scenario, not as instructions that reveal the UI. 3. Add a short pre-test screener and 3 post-task questions. 4. Define the success criteria and what observations would invalidate the design. </task>
Produces a full unmoderated test script with goal-based tasks, screener, and clear success criteria.
Pro tip: Ask ChatGPT to rewrite any task that 'leaks the answer' — leading tasks are the most common reason usability results mislead.
Responsive Behavior Spec
20/30<context> The [FEATURE] screen in [PRODUCT] must work across mobile, tablet, and desktop for [USER]. I need to spec the responsive behavior before prototyping. </context> <task> 1. For each breakpoint (mobile, tablet, desktop), describe the layout and what changes from the previous size. 2. Specify which elements reflow, stack, hide, or collapse into menus. 3. Define touch-target and spacing rules for the smallest breakpoint. 4. Flag the single element most at risk of breaking responsively and propose a fallback. </task>
Specifies layout behavior across breakpoints with reflow rules, touch targets, and at-risk element fallbacks.
Pro tip: Give ChatGPT your actual breakpoint pixel values so the reflow advice maps to your real grid instead of generic ranges.
Design Critique
5 promptsHeuristic Evaluation
21/30<context> I want a structured critique of my [FEATURE] design in [PRODUCT] for [USER]. I will describe the screens and key interactions below: [DESCRIBE DESIGN]. </context> <task> 1. Evaluate the design against Nielsen's 10 usability heuristics. 2. For each heuristic, state whether it is met, partially met, or violated, with a specific reason. 3. Rank the violations by severity (cosmetic, minor, major, catastrophic). 4. Give one concrete fix for each major or catastrophic issue. </task>
Runs a heuristic evaluation with per-principle verdicts, severity ranking, and concrete fixes.
Pro tip: Describe the design in plain words OR paste a screenshot if using GPT-4o vision — ChatGPT critiques visuals directly when you attach the image.
Devil's Advocate Reviewer
22/30<context> I am attached to my current direction for [FEATURE] in [PRODUCT] and want someone to poke holes before I present it. Here is the design and my rationale: [DESCRIBE]. </context> <task> 1. Argue the strongest case AGAINST this design direction. 2. Identify the 3 assumptions it relies on that could be wrong. 3. Describe a realistic scenario where [USER] would be frustrated or confused. 4. Suggest the single change that would most de-risk the design. </task>
Stress-tests your direction by attacking assumptions, surfacing failure scenarios, and naming the top de-risking change.
Pro tip: Tell ChatGPT 'do not soften the critique' — it tends toward politeness, and the harsh version is what saves you in the review.
Accessibility Critique
23/30<context> I need to audit my [FEATURE] design in [PRODUCT] for accessibility before handoff. The design is described here: [DESCRIBE DESIGN AND COLORS]. </context> <task> 1. Check the design against WCAG 2.2 AA for color contrast, focus, target size, and text alternatives. 2. List each likely violation with the specific guideline number. 3. Recommend a fix for each, with concrete values where relevant (e.g. contrast ratio targets). 4. Flag anything that needs real assistive-tech testing rather than a desk review. </task>
Audits a design against WCAG 2.2 AA, listing violations by guideline with concrete fixes.
Pro tip: Paste your hex color pairs into the context so ChatGPT can estimate contrast ratios instead of giving you generic advice.
Copy and Microcopy Critique
24/30<context> I want to tighten the UX writing in [FEATURE] of [PRODUCT] for [USER]. Here is all the current copy: [PASTE COPY]. </context> <task> 1. Critique each string for clarity, length, tone, and action-orientation. 2. Rewrite each weak string, keeping our voice consistent. 3. Flag any jargon, hedging, or vague CTAs and replace them. 4. Ensure button labels describe the outcome, not the mechanism. </task>
Critiques and rewrites every UX string for clarity and voice, fixing vague CTAs and jargon.
Pro tip: Provide 2-3 examples of your existing on-brand copy first so ChatGPT matches voice instead of defaulting to generic SaaS tone.
Design Decision Defender
25/30<context> A stakeholder is pushing back on a design decision I made for [FEATURE] in [PRODUCT]. My decision: [DECISION]. Their objection: [OBJECTION]. </context> <task> 1. Steelman the stakeholder's objection so I fully understand it. 2. Lay out the strongest design rationale for my decision, grounded in user needs and principles. 3. Identify any valid point in their objection I should incorporate. 4. Draft a calm, evidence-based response I can say in the meeting. </task>
Steelmans the objection, builds your rationale, and drafts a calm evidence-based meeting response.
Pro tip: Ask for the response in spoken language, not written — ChatGPT defaults to memo prose that sounds stilted said aloud.
Go from copy-pasting to actually mastering AI.
AI Academy: 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ other tools. New tutorials added every week.
Developer Handoff
5 promptsHandoff Spec Generator
26/30<context> I am handing off [FEATURE] of [PRODUCT] to engineering. I need a complete spec document so developers can build without back-and-forth. </context> <task> 1. Generate a handoff spec with sections: Overview, User flow, Screens and states, Component list, Interaction behavior, Edge cases, and Acceptance criteria. 2. Write acceptance criteria in clear given/when/then format. 3. Note any dependencies, API needs, or open questions for eng. 4. Keep it scannable with headings and bullets, not walls of text. </task>
Produces a complete, scannable handoff spec with given/when/then acceptance criteria and open questions.
Pro tip: Ask ChatGPT to format the output in Markdown so you can paste it directly into Notion, Linear, or your ticketing tool.
Design Token Documenter
27/30<context> I need to document the design tokens used in [FEATURE] of [PRODUCT] so engineering applies them consistently. My values are: [PASTE COLORS, SPACING, TYPE]. </context> <task> 1. Organize my raw values into a token structure: color, spacing, typography, radius, shadow. 2. Propose semantic token names (e.g. color-surface-primary) mapped to my raw values. 3. Flag any inconsistent or one-off values that should be normalized. 4. Output the tokens in a clean table ready for the design system doc. </task>
Structures raw style values into semantic design tokens and flags inconsistencies for normalization.
Pro tip: Ask for the tokens as a JSON object too — many design-to-code tools import token JSON directly, saving manual entry.
Acceptance Criteria Writer
28/30<context> The [FEATURE] design in [PRODUCT] is ready and I need precise acceptance criteria so QA and eng know exactly what 'done' means for [USER]. </context> <task> 1. Write acceptance criteria for every screen and key interaction in given/when/then format. 2. Include criteria for edge cases, error states, and empty states. 3. Add accessibility acceptance criteria (keyboard, focus, contrast). 4. Flag any criterion that is ambiguous and needs a product decision before build. </task>
Generates thorough given/when/then acceptance criteria covering edge cases, accessibility, and ambiguities.
Pro tip: Tell ChatGPT to 'mark any criterion you could not fully specify from my input' so gaps surface before QA finds them.
Eng Question Pre-empter
29/30<context> Before I hand off [FEATURE] of [PRODUCT] to engineering, I want to anticipate the questions developers will ask so I can answer them in the spec. </context> <task> 1. List the 10 most likely questions a frontend developer would ask about this design. 2. Group them by theme: data, states, interaction, responsive, and edge cases. 3. For each, note what I should add to the spec to pre-empt it. 4. Highlight the 3 questions that, if unanswered, would most likely block the build. </task>
Predicts the developer questions a handoff will trigger and tells you what to add to pre-empt blockers.
Pro tip: Paste your draft spec and ask ChatGPT 'which of these 10 are already answered?' so you only fill the genuine gaps.
Implementation Review Checklist
30/30<context> Engineering has built [FEATURE] for [PRODUCT] from my designs. I need to do a design QA review of the implementation against the spec for [USER]. </context> <task> 1. Generate a design QA checklist covering visual fidelity, spacing, typography, states, interactions, and responsive behavior. 2. Add a section for accessibility checks and one for motion/animation timing. 3. Phrase each item as a yes/no check I can run quickly against the live build. 4. Recommend a severity label for each item so I can triage what blocks release. </task>
Produces a yes/no design QA checklist with severity labels to triage implementation issues before release.
Pro tip: Ask ChatGPT to order the checklist by where bugs most often hide — states and responsive edges — so your review catches the worst first.
Frequently Asked Questions
Prompts are the starting line. Tutorials are the finish.
A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.
7-day free trial. Cancel anytime.