A framework for intentional AI use.
Not all design work should be AI-accelerated. Some parts of the process need to remain deeply human, not for romantic reasons about craft, but because that’s where the actual value lives.
I’ve found it helpful to think about design work across two dimensions: the level of creative judgment required, and whether speed and volume matter.
- Work requiring minimal judgment but high volume: generating design variations, resizing assets, creating component states. This is where AI excels. Let it handle the repetitive execution while humans set the direction and review for quality.
- Work requiring exploration and refinement: wireframe layouts, visual design directions, UX flows. This is where AI and humans collaborate best. AI accelerates getting to 70%, humans provide the strategic thinking and refinement that makes it excellent. (A case study by a Czech company shows that their proprietary AI agent allowed them to accelerate the design, testing and creation of MVPs, with a reported 40% efficiency gain in graphic design and a 200% increase in content creation speed.)
- Work requiring original creative judgment: campaign creative, brand systems, strategy, storytelling. This needs to remain human-first. AI can support research and inspiration, but humans must own creative thinking. (Some research indicates that AI-powered brainstorming is actually more powerful than human-only creativity alone.)
- Work involving product accuracy, legal requirements or sensitive content: This should have minimal or no AI involvement due to risk.
The framework isn’t about rules, but about being deliberate. Asking on every project: Where does AI create space for better thinking, and where does it just create distance from the work?
Protecting what makes us human.
When a designer spends hours exploring solutions to a complex problem, they’re building intuition about what works and why, developing judgment through iteration, learning to see patterns that inform future decisions.
When AI generates those same explorations instantly, the designer gets the output but misses important processes. This can be more harmful especially to younger talent in the industry.
This matters. The best design work comes from intimacy with the problem. You can’t have intimacy if you’re just prompting from 10,000 feet.
The solution isn’t to avoid AI. It’s to be intentional about preserving the hands-on craft and creative exploration that makes designers better—and that produces genuinely innovative work.
- Early conceptual exploration should remain hands-on.Sketching ideas, whiteboarding user flows, roughing out approaches—this is where designers develop a point of view and understand problems deeply. AI can help execute those concepts quickly for testing, but not before designers have defined what to explore.
- Craft refinement needs human attention. The final 20% that makes work feel premium rather than generic: the subtle spacing adjustment, the unexpected color choice, the typography that feels exactly right. This requires a designer’s eye and judgment.
- Creative ideation benefits from constraint. When you’re forced to sketch five concepts by hand, you think differently than when you prompt AI to generate fifty options. The constraint forces creativity. The manual process surfaces ideas that may not emerge from prompting.
- Novel problems need human exploration. When facing something truly new (e.g., a user need that hasn’t been solved before, a brand expression that needs to break from category norms) designers need space to explore without the pattern-matching tendencies of AI steering them toward familiar solutions.
If teams over-index on AI and lose touch with hands-on craft—if everything becomes prompting and curation—the work risks feeling generic. AI’s speed could push teams to build first and think later, prioritizing function over the conceptual exploration that creates meaningful innovation. I’m not saying that we’ll go 100% all in on AI, but we have to not only consider who we are designing for, but also who we are designing with.
The best integration of AI preserves and enhances human creativity, handling repetitive execution so design teams can bond, debate, discuss, explore, ultimately focusing more energy on strategic and creative work that humans do better.
What becomes possible.
On a recent project designing an AI-powered real-estate application, we used Figma Make to generate sacrificial prototypes. Traditionally when conducting initial design research we would only have time and resources to produce sketches or very rough prototypes. With AI we are able to make much deeper, functional prototypes.
With AI we were able to create more time for research and validation:
- By sharing the sacrificial prototypes with users we were able to sharpen the requirements and shape the product to better fit their needs.
- We uncovered nuances with the user’s pain points that allowed us to design confidently and not second guess certain features.
- We spent time as a team debating strategic direction instead of rushing to execution.
AI didn’t make us faster—it made us more thorough. Overall we saved at least 2 weeks of project time that created space for this adhoc research.
Leading through the transition.
If you lead a design team, you’re likely navigating this transition: designers are exploring AI tools, while clients are asking questions about its use.
Thoughtful integration means:
- Being explicit about when and how AI should be used, rather than deciding case-by-case on every project. - Protecting the parts of the process that need to remain deeply human—the hands-on craft, creative exploration, and strategic thinking that create value.
- Reinvesting the time AI saves into deeper research, more iteration, and better strategic thinking rather than just shipping more volume.
- Ensuring teams maintain hands-on craft skills and creative confidence, not just AI prompting abilities.
- The teams that thrive will be the ones who know exactly which parts of the process to protect as deeply human, and which to accelerate with AI.
The real opportunity.
The promise of AI in design isn’t that we can ship more work with fewer people.
It’s that we can design more thoughtfully: more thoroughly researched, more extensively explored, more rigorously tested, more genuinely centered on user needs.
Like most designers I share apprehension about certain aspects of AI bleeding into our work, but as it matures and we augment our process I am hopeful and excited about how it can empower us to create space and opportunity.
The question for design leaders right now is: What will you do with the space AI creates?
