If you have a design system in Figma and want to use for crafting design in Claude Code, this tutorial is for you: https://youtu.be/mqGvJ_g2GME My approach is to create two Claude Skills - one for design system tokens (styles) and another for UI components from your design system. And only after that craft actual design:

I’m seeing more and more designers proudly stating “No AI used” for the fine UI design, graphics, and motion effects they craft. At first, I assumed that these people are against the technology itself. But after looking deeper than that, most of "No AI used" designers have a different perspective on how products should be crafted. I've shared my thoughts on No AI used in this article: https://medium.com/ux-planet/no-ai-used-4932970c7db8?sk=ff9f862136ca1195684dc3614bec4f7a What do you think?

AI era has caused significant disruption across many fields, but product design feels like one of its primary targets. The rapid improvement of AI tools, combined with constant competition among major tech players, has led to the release of AI tools that allow almost anyone to create their own products in no time. This shift has had a major (largely negative) impact on professionals who were responsible for digital product design in the pre-AI era: UX designers, UI designers, product designers, accessibility specialists, and others. A lot of people approached me asking how to adapt to this new reality, so I decided to craft a set of 5 rules I personally practice: 1️⃣ Invest in improving your critical thinking and taste Taste and critical thinking are what will separate average product creators from exceptional ones in the AI era. Your workflow will increasingly revolve around analyzing AI-generated output and aligning it with user needs and business goals. - Critical thinking helps you articulate what you want - Taste helps you recognize and avoid mediocre output. The goal of using AI isn’t to produce more design in less time; it’s to produce better design in less time. 2️⃣ Use AI tools as your companion in every step of the product design process Actual AI tools don’t really matter; what matters is the person who uses the tool and how they use the tool. AI shouldn’t be treated as a one-off tool; it should be integrated into every stage of your workflow because the tools are just multipliers of your skills and abilities. From early research and ideation to prototyping, testing, and implementation, AI can support and accelerate your work. The key is to shift your mindset: instead of “using AI when needed,” think of AI as a constant collaborator. 3️⃣ Forget about wireframing Time is your most valuable resource. Don’t waste it creating static design artifacts, especially wireframes. 4️⃣ Don’t be obsessed over crafting “perfect” prompts Focus on the task you’re trying to accomplish. One approach that helps me avoid overthinking prompts is using voice input. When I use voice in tools like Claude Code, I don’t worry about perfect phrasing; I focus on clearly expressing what I want to achieve. For me, voice is one of the most effective ways to interact with AI. It naturally shifts your attention away from wording and toward intent, which is what actually matters.

If you want real leverage when using Claude Code for complex product design tasks, do the following two things every time: 1. Plan your project → Ultraplan 2. Review your code before merging to master → Ultrareview 🧠 UltraPlan (plan before you build anything complex) Vague input leads to generic output. Thats why when using Claude Code, don’t jump straight into generation. Instead, start with a detailed and clear plan: Use Claude Code in Ultraplan mode to: ✅ break down features ✅ define architecture ✅ map UI → logic → data This helps reduce ambiguity before writing a single line of code 📕 Guide to UltraPlan mode: https://medium.com/ux-planet/the-best-way-to-plan-work-with-claude-code-58e336a3b0ba?sk=129d17578c8e71b0f8d76912719a13db 🔍 UltraReview (review like a team of senior engineers) In UltraReview mode, Claude runs a few AI agents in parallel to review source code and catch hidden bugs, poor structure, and weak solutions. In my experience, Ultrareview is vital for project milestones (i.e., when you release a new feature and want to make sure it doesn’t have bugs). 📕Guide to UltraReview: https://medium.com/ux-planet/how-to-improve-code-quality-with-claude-code-4e529ded8b4e?sk=b60321c0a810077851260f2ef9fe515e

https://medium.com/ux-planet/claude-design-figma-killer-or-just-another-design-tool-82f7726693ca?sk=5ff56983994561a73dc146dffa1fc62e Figma killer or just another AI design tool? What do you think?