Designing with and around AI
Last updated: June 2026
Designing with and around AI means using it as a powerful collaborator for speed and exploration while maintaining clear boundaries on what you hand off to it, never delegating final judgment, taste, or responsibility.
The Principle
AI is an incredible amplifier for certain parts of design and development work: generating variations, synthesizing research, writing first drafts, exploring layouts, and handling repetitive tasks. It is fast, tireless, and often surprisingly creative within its training distribution. However, it lacks taste, context awareness, ethical judgment, and deep understanding of your specific users or brand.
The practical approach is to use AI for divergence and heavy lifting, then bring strong human judgment for convergence and final decisions. Honest boundaries include: never let AI write final client communication, make final design decisions, or handle high-stakes accuracy without verification. Use it to explore, iterate, and remove drudgery — not to replace your critical thinking.
In my own practice, I went through the typical honeymoon phase of asking AI to do almost everything. Some outputs were brilliant; many required heavy editing or were subtly wrong in tone or logic. Learning to use it as a very capable intern rather than a replacement designer was liberating. It now saves me hours on exploration while I retain control over direction, quality, and final execution.
Why It Matters for Design & Building
AI is moving fast, and the designers who use it effectively will have a significant advantage in speed and output. However, those who delegate too much risk producing generic, ungrounded work and losing their own voice and judgment over time.
As a Design Engineer, I now have clear rules: I use AI for initial research summaries, copy variations, layout explorations, and code snippets. I never use it for final microcopy, brand-sensitive decisions, or client-facing deliverables without heavy editing. In one recent project, AI helped me generate 30 layout variations in minutes, but I made the final choices and refinements myself. The result was faster exploration without sacrificing quality or personal touch.
This balance is essential for calm technology and thoughtful building. AI should reduce stress and busywork, not create new sources of doubt or ethical compromise. The honest practice is to use AI to extend your capabilities, not to outsource your responsibility.
Real-World Examples
My own current workflow is the clearest example. I use AI daily for brainstorming copy variations, summarizing competitor research, and generating initial code structures. I then refine everything myself. This has roughly doubled my exploration speed while keeping the final output distinctly mine.
Many junior designers I’ve mentored fall into the trap of letting AI write entire flows or interfaces. The results often look polished but feel generic and lack strategic depth. Clients notice the difference immediately.
A client project involving a wellness app offered a mixed case. We used AI heavily for generating mood board ideas and initial copy options. However, we kept all final visual direction, tone of voice, and user flow decisions human-led. The project moved faster than previous ones, but the end product still felt intentional and human-centered rather than AI-generated.
References
- Case, A. (2015). Calm Technology. O'Reilly Media.
- Jarvis, P. (2019). Company of One. Houghton Mifflin Harcourt.
- Basecamp. (2019). Shape Up: Stop Running in Circles and Ship Work that Matters. Basecamp.
- Budiu, R. (2023). "Explainable AI in Chat Interfaces." Nielsen Norman Group. nngroup.com
New entries are published every 2–3 weeks.
Follow along on X or LinkedIn to get notified.