The Art of Prompting—How Architects, Product Designers, and Artists Shape AI Creativity
Nithya Subramaniam
Senior Product Designer | Investor | Experience Design | Design Architect
AI is not a magic wand. It’s more like a really eager intern—one who follows instructions too literally and sometimes brings back results that make you go, huh?
But here’s the thing: prompting AI is a skill. And the best prompters aren’t just engineers—they’re architects, product designers, and artists. Why? Because they already think in systems, constraints, and creativity.
This week’s Triple Take breaks down how prompting AI is like designing a building, creating a product, or making art. Let’s dive in.
1?? Architecture: Designing the Blueprint for AI Creativity
Imagine giving an architect a vague brief:
?? “Make me a house.”
You’ll get a house. But is it a minimalist cube? A Victorian mansion? A treehouse with a slide? Without the right level of context and constraints, you get whatever the architect feels like sketching that day.
AI works the same way. A prompt like “Generate a futuristic city” will spit out a generic, Blade Runner-esque skyline. But great prompters think like architects:
? Define the foundation (Materials, era, density?)
? Set constraints (Climate, sustainability goals?)
? Detail the user experience (Is it for pedestrians, AI-driven cars, or underwater civilizations?)
?? Example: Architects are now using AI to model adaptive urban spaces that respond to climate change in real-time. Instead of designing static blueprints, AI-generated cities evolve dynamically—kind of like a SimCity where buildings adapt based on real-world weather data.
?? Takeaway:
AI isn’t designing the future—we are. AI is just executing the blueprint.
2?? Product Design: Prompting as UX for AI
If architects think in space, product designers think in flows. And guess what? Prompting AI is just like designing a great user experience. Ever been frustrated by a chatbot that doesn’t understand what you want? That’s bad UX. AI models work the same way—if you don’t structure your input correctly, you get garbage output.
?? Bad AI prompt: “Write a marketing copy for a new app.”
?? Good AI prompt: “Write a 3-sentence marketing hook for a budgeting app targeting Gen Z, using playful and relatable language.”
?? Example: We introduced AI-powered risk control suggestions while I was in TrustCloud (Yes, let's return to an area of expertise ??), users needed accurate, relevant controls—not a laundry list of random suggestions. The secret sauce? Structuring AI inputs like a well-crafted UX flow:
? Who was this for? (Compliance officers at SaaS startups)
? What’s the context? (They need to mitigate risk but don’t want generic recommendations)
? What’s the ideal outcome? (Actionable, high-confidence suggestions that reduce time spent searching for controls)
?? Takeaway:
AI’s biggest problem isn’t intelligence—it’s usability. Great product design makes AI work for humans, not the other way around.
3?? Art: The Dance Between Chaos and Control
Artists know this: true creativity lives between structure and chaos. AI can generate a thousand variations of an idea, but it can’t decide what’s good. That’s where human intuition comes in.
Think about Bharatanatyam, the classical Indian dance form. There’s rigid structure—mudras (hand gestures), rhythm, form. But within that, the dancer brings expression and storytelling. AI can help generate Bharatanatyam-inspired movements, but can it breathe emotion into them? Not yet.
?? Example: AI-generated art, music, and even dance choreography are exploding. But the most successful AI artists aren’t the ones who let AI take over—it’s the ones who curate, remix, and reinterpret.
?? Takeaway
AI is a tool. The best creators are the ones who guide it with intention, structure, and an artist’s touch.
?? Bonus Content: Prompt Like a Pro!
Want to get better at prompting AI? Here are three power moves and simple examples to take your AI outputs from meh to mind-blowing:
1?? The Director’s Cut: Instead of “Generate a sci-fi city,” try:
? “Generate a cyberpunk city with neon-lit skyscrapers, bustling pedestrian streets, and automated food carts. The city should feel alive, like a mix of Tokyo and Blade Runner.”
? ?? Pro Tip: Treat AI like a film director—set the scene, tone, and emotion.
2?? The Iterative Genius: Never settle for the first response. If AI gives you something basic, refine the input:
? “This is too generic. Make it sound more poetic.”
? “Reword this for a 10-year-old audience.”
? ?? Pro Tip: The best AI results come from layered refinements.
3?? The Unexpected Twist: AI thrives on surprise. Throw in an unusual constraint:
? “Generate a Renaissance-style oil painting of a robot contemplating its own mortality.”
? ?? Pro Tip: AI gets creative when you force two worlds to collide.
Try these out and tell me what you discover! ??
Closing Thoughts: So, What’s the Prompt for the Future?
AI won’t replace architects, product designers, or artists. But the ones who master prompting, structuring, and refining AI output? They’ll lead the charge.
This week, let's try this:
?? Architects: Prompt MidJourney to design a liveable future city—see if it understands the human experience.
?? Product designers: Treat AI like a new user—does it navigate your UX flows well, or does it get lost?
?? Artists: Push AI-generated art past randomness—can it tell a story the way you do?
At the end of the day, AI is just pixels and patterns. We bring the meaning.
Until next time—keep prompting with intention! ??AI Prompt
Keep a lookout for my next post on Medium where I cover my learnings from the books and other places.
Check out other recent posts of mine -
The Rebel Architect Who Bent the World to her Curves - was published in the top sweet picks by editors from sweet publications.
Toddlers, Toast, and Task Flows: How Daily Chaos Inspires Great Design - that was published in Practice in Public publication
Award Winning Quality Engineer specializing in Agile Project Management and Jira
3 天前That's a cool image header!