Andrew “Boz” Bosworth's 10 Rules for Navigating the Next Design Paradigm | Figma Blog

Andrew “Boz” Bosworth argues that designers entering the next era of AI and spatial computing must challenge inherited assumptions rather than merely improve existing interfaces. His approach begins with real human problems, favors experimentation and intuition, and treats interactions as complete systems. The goal is to create technology that understands users and ultimately becomes nearly invisible.

Start with real human problems

• Identify a specific person with a genuine problem.
• Let users determine whether a product is useful through their behavior.
• Avoid prioritizing abstract ideas over practical human needs.

Question inherited design assumptions

• Recognize that familiar constraints and interaction patterns are man-made paradigms.
• Ask whether the current approach is actually appropriate.
• Treat seemingly fixed limitations as potentially changeable.

Reimagine the interaction paradigm

• Current computing often forces users to translate simple intentions into complicated sequences of apps and services.
• After decades of desktop and mobile conventions, designers should reconsider the entire model of interaction.
• New technology should simplify intentions rather than expose underlying complexity.

Distinguish invention from optimization

• “Zero-to-one” invention happens without established customers or constraints.
• Optimization improves and refines an existing product with audience feedback.
• AI and spatial interfaces provide greenfield opportunities where old assumptions may not apply.

Use taste and intuition to choose the right direction

• Product development resembles climbing through a difficult problem space.
• Intuition helps teams choose promising “terrain” before investing in execution.
• Good judgment does not remove the hard work, but it improves the odds of pursuing the right problem.

Prototype aggressively

• Build rough, unconventional prototypes to test whether an idea is worth pursuing.
• Meta has used crude physical setups, including tracked hats and mesh-walled rooms, to explore spatial computing.
• Direct experimentation reveals possibilities that discussion alone cannot.

Design the complete system

• Spatial products cannot be designed by changing one isolated component.
• Gestures, visual/audio/haptic feedback, and resulting functionality must evolve together.
• Iteration should happen across the whole interaction system.

Give tools an appropriate theory of mind

• Future tools should understand users’ intentions and goals.
• They need enough agency to assist meaningfully, but not so much that they become intrusive or unpredictable.
• Effective assistance depends on balancing automation with user control.

Treat products and people as works in progress

• Products should be viewed as successive versions rather than finished objects.
• Bugs and improvements are a normal part of continued development.
• Designers should apply the same iterative mindset to their own growth.

Make interfaces disappear

• Interfaces should be as seamless and minimal as possible.
• Their value lies in enabling an experience, not in drawing attention to themselves.
• The ideal interaction removes unnecessary friction between a person and their intention.

Bosworth’s practical recommendation is to stay close to human needs, challenge conventional assumptions, prototype early, and design the entire experience. In emerging fields such as AI and spatial computing, success depends less on polishing familiar interfaces than on inventing simpler, more natural ways for people to accomplish things.