Articles
Thoughts on design + crafting delightful digital products

AI-Ready Design Systems
An AI-ready design system is the source of truth AI coding tools read every session: tokens, specs, and docs, not a new component library to install.
Read the series
Part 110 min readAI-ready design systems: structure tokens, specs, and docs so AI stays on-brand
Your design system can be an AI-ready source of truth when you structure tokens and specs so coding agents load your files instead of guessing from screenshots.
Part 211 min readHow to structure a design system in five tiers (tokens → skills)
Build a five-tier design system structure before AI-generated UI scales. An AI-ready folder stack for SaaS product teams using Cursor and Claude Code.
Part 310 min readHow to audit a design system before you rebuild (or feed it to AI)
Audit your design system for drift and dead UI before you scale AI generation. Checklist, and when to bring a partner.
Part 49 min readTailwind v4 semantic design tokens + WCAG contrast for Next.js (2026)
Semantic design tokens in Tailwind v4 for Next.js. Two layers, one source of truth, so accessible color contrast pairs stay intentional and WCAG-minded.
Part 510 min readComponent specifications AI tools can implement without guessing
The exact format for component specs for AI: YAML frontmatter, nine sections, four status values, and the failure modes that let drift creep in.
Part 68 min readWhen design system tokens and live code disagree: how to reconcile
Two files claim one token. That's token drift. Live code usually wins. A four-category rubric sorts Tailwind vs tokens.css, and often finds a WCAG bug.
Part 711 min readCreate a Claude Code skill for your design system (tokens + Taste Profile)
Stop pasting design rules into Claude. Ship a Claude Code skill for your design system so every session loads the same tokens, voice, and brand rules.

The Judgment Layer
As AI collapses the cost of execution, the most valuable thing a product team can have isn't faster output, it's clearer judgment about what to build and why. Crafting this first helps you move at the speed of thought when executing.
Read the series
Part 14 min readThe Execution Collapse
Products that succeed in planning often fail in implementation. Explore the causes of the Execution Collapse and practical strategies to avoid it on your team.
Part 24 min readThe Three Roles Are Becoming One (And That's a Problem)
AI has compressed the skill floor across design, engineering, and product. Most practitioners can now operate in all three. That's not the real problem.

Branding is your product's personality
Branding has shifted from a marketing tactic to a product decision. When you're building fast, branding feels like something you can postpone. But users don't experience “later.” They experience the product exactly as it shows up today. Brand is a product's personality that makes someone coming back for more or forgetting you, and your product team is responsilble for shaping that experience.
Read the series
Part 16 min readBranding is a product decision, not a marketing layer
Small teams feel the absence of branding most. Lack of clarity turns speed into chaos, you move fast, but not in a straight line. Here's how to fix that.
Part 25 min readDefining product identity early is cheaper than fixing it later
Identity debt compounds like technical debt, but you can't refactor it in a weekend. Defining what your product is trying to be is the cheapest decision.
Part 35 min readHow feature growth quietly separates products from their identity
Features accumulate faster than meaning. Users feel drift as friction long before they can name it. How to recognize misalignment before it compounds.
Part 45 min readConsistency is how unknown products earn trust
Unknown products don't get the benefit of the doubt. Users read signals long before they commit. Consistency isn't polish. It's how trust forms.
Part 55 min readDifferentiation comes from clarity, not novelty
Novelty fades. Clarity compounds. The strongest differentiation for small teams isn't a unique feature, it's being understood quickly.
Part 66 min readWhy speed comes from constraints, not options
The real bottleneck for small teams isn't execution, it's choice. Constraints collapse decisions, turning 'Is this good?' into 'Is this us?'
Part 74 min readWhy brand clarity turns speed into something durable
The real tradeoff isn't speed versus brand, it's speed versus fragility. Rework is the hidden tax of moving without direction. Brand clarity anchors speed.
Part 86 min readScale breaks products that don't know what they are
Growth magnifies whatever already exists. Intuition doesn't scale. Brand is shared context that lets teams operate independently without losing coherence.

The New Craft
As AI tools make professional-looking design accessible to everyone, craft becomes the differentiator. This series explores how real design craft evolves in an AI-driven world, and how to spot the difference between commodity and craft.
Read the series
Part 15 min readEveryone has the same AI tools. That's exactly why craft matters more
AI tools made 'good enough' free. If your competitor can generate the same landing page in ten minutes, you don't have a brand. You have a template.
Part 25 min readDesign isn't the deliverable: It's the hidden work that makes products succeed
AI can generate the visible 10% of design. But the invisible 90% is what makes products actually succeed. This is where craft lives, and where AI can't go.
Part 35 min readStanding out when everyone has the same tools
Which AI startup has the purple gradient and sparkle icon? They all do. When everyone uses the same tools, the brands that survive know who they are.
Part 47 min readMy actual AI workflow after 22 years in design
After 22 years of building products, here's exactly how I use AI: the specific tools, the actual workflows, and where I still do things by hand.
Part 56 min readThe multiplication effect: Why AI-augmented teams beat AI-augmented individuals
One person + AI = more output. Multiple experts + AI = exponentially better outcomes. Solo you touch the surface. With a team, each expert brings real depth.
Part 67 min readThe craft test: Questions that reveal whether you're getting real design or AI commodity
You're done not when there's nothing left to add, but when there's nothing left to take away. As AI floods the market, spotting real craft becomes critical.

Designing for AI
AI products require fundamentally different design thinking. This series covers the patterns, principles, and practices for building AI-powered experiences that users actually trust and enjoy.
Read the series
Part 16 min readWhy AI products fail on UX, not the model
Most AI products fail on UX, not the model: black-box outputs, unclear handoff, weak onboarding, and broken trust. Design that before you tune the model.
Part 24 min readHow to Design User Experiences That Make AI Feel Human
Product design is evolving from interfaces to intelligent interactions. How to create AI experiences that feel trustworthy and natural.
Part 38 min readBuilding trust in AI systems: UX patterns that make outputs believable
Trust in AI products is a design problem: show uncertainty, give control, and recover from mistakes before you tune the model.
Part 47 min readDesigning for AI failures: error states, recovery, and trust when the model is wrong (2026)
AI will fail. Design error states, confidence cascades, graceful degradation, and a human handoff so people still trust the product when the model is wrong.
Part 57 min readHow to Turn AI into a Co-pilot, Not a Black Box
The future isn't bots that replace people. Learn how to design co-pilot experiences where AI becomes a trusted teammate that enhances human decision-making.
Part 69 min readDesigning Conversational AI: Beyond the Chatbot Paradigm
Chatbots are just the beginning. Learn how to design conversational AI that feels natural and helpful through better prompt design and conversation flows.
Part 78 min readAI in Traditional Interfaces: Beyond Chat Bubbles
The future of AI isn't chat windows. Learn how to embed AI suggestions, automation, and intelligence directly into traditional interfaces users already know.
Part 87 min readTesting & Iterating AI Features: Beyond Traditional UX Methods
Traditional usability testing falls short for AI features. Explore specialized testing methods, metrics, and iteration strategies for AI-powered experiences.
Part 98 min readMulti-Modal AI Experiences: Designing Beyond Text
Learn how to design multi-modal AI interfaces that work seamlessly across text, voice, vision, and gesture while maintaining consistency and user control.
Part 108 min readAdvanced AI Patterns: Personalization, Learning, and Adaptive Systems
Explore advanced patterns for AI that learns, personalizes without being creepy, and adapts to individual and team preferences over time.
Part 119 min readImplementing AI Design Systems: Components, Guidelines, and Team Processes
Learn how to build design systems for AI components, establish guidelines for AI behavior, and create team processes that scale AI design effectively.
Part 1210 min readThe Future of Human-AI Collaboration: Emerging Trends and Design Implications
Emerging trends in human-AI collaboration, from autonomous agents to brain-computer interfaces, and what they mean for designers building AI experiences.