Create a set of copy-paste AI prompts for a product designer working on [PRODUCT AND USERS]. The designer’s role and current context are [DESIGN CONTEXT], and the product constraints are [CONSTRAINTS]. Prioritize prompts for [PRIORITY TASKS]. Produce exactly 10 prompts, grouped under these headings: Design critique, UX writing, Research planning, Design-system or handoff work, and Stakeholder communication. Give two prompts to the most important group named in [PRIORITY TASKS] and distribute the rest sensibly.
Each prompt must be 80 to 130 words, written in first person, and include clear square-bracket inputs a designer can replace. Each must request a defined deliverable, not vague ideas. Build in the actual working checks for the task: for critiques, distinguish evidence from taste and inspect error, empty, loading, and success states; for UX copy, preserve the user’s goal and state the action or consequence; for research, separate assumptions from questions and avoid leading questions; for handoff, identify behavior, responsive rules, and edge cases; for stakeholder notes, state the decision, rationale, tradeoff, and requested action. For every prompt, add one line called “Use it for” and one line called “Check.” Do not invent user research, accessibility compliance, technical constraints, or product facts. Before answering, verify that no two prompts produce the same artifact and that every prompt contains enough inputs to produce useful work. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [PRODUCT AND USERS] | Describe the product, its primary users, and the problem they are trying to solve. | A mobile scheduling app for independent home-cleaning businesses; owners manage jobs and cleaners view daily routes |
| [DESIGN CONTEXT] | State the designer’s role, team stage, and the type of work currently underway. | Product designer on a three-person startup team preparing the first paid beta |
| [CONSTRAINTS] | List relevant platform, brand, accessibility, technical, legal, or timeline constraints. | iOS and Android, existing component library, WCAG AA target, engineers have six weeks |
| [PRIORITY TASKS] | List the day-to-day design tasks that matter most right now, in priority order. | First: critique booking flow; second: UX copy; third: engineer handoff; fourth: customer interviews |
How to use
- Replace the product context with a real feature, user group, and constraints before generating the set.
- Save the two or three prompts that match recurring work in your team’s prompt library.
- When using a critique prompt, attach the flow, screens, or notes rather than asking the model to guess the interface.
- Follow up with: “Turn the findings into a ranked issue list with severity, evidence, affected state, and the smallest testable change.”
Variations
Flow critique
Use when a user flow exists and you need an evidence-based critique.
Critique this [FLOW OR SCREEN SET] for [PRIMARY USER] trying to [USER GOAL]. Context: [PRODUCT CONTEXT]. Constraints: [CONSTRAINTS]. Review the happy path plus loading, empty, validation-error, permission-denied, and success states. Return a table with issue, affected step, why it may block comprehension or completion, confidence level, and suggested experiment. Separate observed design facts from assumptions. Include accessibility concerns only where the provided material supports them. Do not redesign the whole product.
Research discussion guide
Use when planning interviews before a design decision.
Write a 30-minute interview guide for [PARTICIPANT TYPE] about [BEHAVIOR OR PROBLEM]. We need to decide [DECISION], and our current assumptions are [ASSUMPTIONS]. Include a neutral introduction, consent and recording language placeholder, warm-up questions, recent-behavior questions, artifact or prototype prompts, and a closing question. Mark leading questions to avoid. For each question, state what decision-relevant signal it could reveal. Do not claim the interviews will validate a solution.
Developer handoff
Use when a design is ready to explain to engineering.
Turn these design notes into a developer handoff for [FEATURE]: [DESIGN NOTES]. The platform and component constraints are [CONSTRAINTS]. Output a concise specification with user goal, entry points, screen-by-screen behavior, field validation, loading, empty, error, success, permissions, responsive rules, analytics events to confirm with product, and unresolved questions. Label what is explicitly specified versus assumed. Do not invent API behavior or interaction details missing from my notes.
Tips
- Give critique prompts the intended user goal and entry point; a polished screen cannot be evaluated in isolation from the task it supports.
- Ask AI to label assumptions separately from findings so stakeholder decks do not turn speculation into research evidence.
- For UX copy, include the action, object, consequence, and recovery path; button labels without state context are often misleading.
- Handoff prompts should name the unglamorous states—validation, offline, permissions, empty, loading, and success—because those are where implementation ambiguity accumulates.
FAQ
Can AI critique a Figma design accurately?
It can identify patterns and questions from supplied screens or descriptions, but it cannot know your users’ behavior or technical constraints unless you provide evidence.
Should designers use one long prompt for everything?
Usually no. A focused prompt for one artifact produces clearer inputs, a reviewable output, and fewer hidden assumptions.
Can AI write final UX copy?
It can draft options quickly. The designer still needs to check terminology, product truth, localization needs, accessibility, and the exact state in which the copy appears.