Crafting effective text prompts for Design: Let’s get specific
Federica Monticelli
on
Hi, I’m Fed!
I’m a UX/UI Designer here at Box UK, and if there’s one thing that gets me genuinely excited (okay, maybe a little bit obsessed), it’s crafting truly brilliant digital experiences. The kind that just works.
Recently, I’ve been undertaking a lot of research, including the excellent in-depth “AI for Designers” course from the Interaction Design Foundation, led by the incredibly insightful Ioana Teleanu. Alongside this, I’ve had a front-row seat to the exciting research and development we’re doing here at Box UK. All of this has completely transformed how I see technology’s role in shaping businesses and, of course, design!
So, if you’re curious about how AI can be a powerful ally, especially when we sprinkle some human design magic on top, I’d love to share with you what I’ve learned.
In this new series, “AI for Decision-Makers,” I’ll be exploring how Artificial Intelligence can genuinely empower organisations like yours to design smarter, move faster, and build with even greater impact. Whether you’re a business leader, a hands-on product owner, or simply curious about the vast potential of AI, my aim is to break down complex ideas into practical, human-centred insights.
Expect a sprinkle of excitement, a big dose of strategic value, and absolutely no baffling jargon!
Begin your AI journey with us
Whether you’re taking your first steps or looking to advance your existing AI strategy, our team is here to provide the knowledge and support you need.
Welcome back to my “AI for Decision-Makers” series, your practical guide to understanding how AI can support real-world design work. We’ve explored collaborating with AI and why prompt engineering matters—now it’s time to get practical.
Yes, you can write prompts that generate:
UX microcopy
Content hierarchies
Research summaries
Design ideas
Interaction flows …and much more!
Like any design process, the magic is in the detail. Let’s walk through how to craft effective text prompts for real-world design tasks, including tips, templates, and examples you can adapt.
What makes a “good” prompt?
A strong text prompt for design is:
Clear: The AI knows exactly what to do.
Contextual: It understands your user or product situation.
Constrained: You set rules, tone, format, or style.
Purposeful: It produces something useful in the design process.
Most ineffective prompts fail because they are vague. Great prompts function like mini creative briefs: short, focused, and user-centered.
Use this prompt formula
Here’s a reliable formula you can adjust for almost any design-related task:
“Write [type of content] for [platform/context]. Tone: [tone]. Format: [format]. Limit: [word/character count]. Audience: [who].”
This approach makes your request easier for AI to understand and gives you more control over the output.
Prompt templates for designers
1. UX microcopy
“Write a tooltip for a disabled 'Submit' button on a checkout form. Tone: helpful and brief. Limit: 15 words. Audience: general consumers, aged 30-50.”
Why it works: Context, intent, tone, length, and audience are all defined.
2. Empty states
“Generate 3 friendly empty-state messages for a project dashboard with no content yet. Include subtle encouragement and 1 call to action. Max 30 words.”
Why it works: Provides variety while setting boundaries and UX goals.
3. Error messages
“Write 2 error messages for a form field where the user entered an invalid email. Use plain language. Keep each under 80 characters.”
Why it works: Functional context, language, and space constraints are clear.
4. Onboarding or first-time use
“Create 3 welcome messages for a mobile app that helps users track their nutrition. Tone: friendly, confident. Audience: busy professionals. Max 20 words each.”
Why it works: Device, purpose, audience, tone, and variation are specified.
5. Accessibility guidance
“Rewrite this button label to be accessible for screen readers. Keep it descriptive, action-oriented, and under 4 words. Original label: ‘Click here’.”
Why it works: Clear input, task, and accessibility constraints ensure inclusivity.
Advanced prompt techniques
Want better results? Try these strategies:
1. Provide examples first
“Here are 2 examples of good empty states. Now create a new one for [context].”
AI, like humans, learns better with examples.
2. Chain prompts together
Break complex requests into smaller steps:
“List 3 user goals when using a calendar app.”
“Now write an onboarding message for each goal.”
This keeps your process structured and aligned with UX flows.
3. Use roleplay
“Pretend you're a UX writer at a fintech company. Write copy for a security warning on a login page.”
Encourages AI to adopt relevant language and tone.
4. Define a voice or style guide
“Write this in the voice of a modern, human-centered brand: clear, concise, warm, with no jargon.”
With 6 years of experience in digital agencies, Federica brings a unique blend of Italian and British UX/UI design, grounded in a strong academic foundation in user experience, psychology, and front-end basics. Her passion for user-centred design and Agile methods shines through in her commitment to creating intuitive, impactful prototypes through iterative design.
Want to chat more?
Get in touch to book your free consultation with our expert team.
Read more: Prompt engineering
Read more: AI as Our Design Partner
Read more: Generative AI: Prepare to be amazed