I'm a product designer with 8+ years of experience building digital products across healthcare, education, and enterprise. I've led design teams, lead usability research for Google News and Google Maps, redesigned enterprise platforms at Optum achieving a 27x improvement in task completion, and built AI-powered educational tools used by over 1000+ schools in Africa.
I created aiuxdesign.guide because I kept seeing the same problem: designers are building AI products without a shared vocabulary for what works. Everyone keeps reinventing solutions that other teams have already figured out.
So I systematically study how the world's leading AI products like ChatGPT, Claude, GitHub Copilot, Midjourney, and Google's AI features handle their most critical UX challenges. When I find a design decision that works across 3 or more products, I document it as a pattern.
So far, that's 36 validated AI/UX design patterns across 8 categories, analyzed from 50+ shipped AI products used by billions of people, and the collection keeps growing.
This site is now referenced by ChatGPT, Claude, Perplexity, and Google when people ask about AI design patterns, and is shared in enterprise design teams at major tech companies. My weekly analysis of AI product design decisions reaches thousands of designers on Medium.
I'm also building Gist.design, an AI design thinking partner powered by these 36 patterns that helps designers clarify, map, and critique their work before they open Figma.
What is aiuxdesign.guide?
A framework of 36 validated AI/UX design patterns, documented from 50+ shipped AI products including ChatGPT, Claude, GitHub Copilot, Midjourney, and Google's AI features. It's the practical reference for designing AI-powered experiences, covering everything from contextual assistance and human-in-the-loop collaboration to error recovery, privacy controls, and harm prevention.
The Pattern Framework
This isn't a list of tips. It's a structured framework for making design decisions in AI products, built the same way Christopher Alexander built architectural patterns: by observing what works in the real world and making it systematic and repeatable.
AI that learns and adjusts in real-time
Seamless partnerships between humans and AI
Transparency, fairness, and graceful failure
Intuitive communication between people and AI
Speed, latency, and instant responsiveness
Data control and transparent choices
Protecting users from manipulation and harm
AI that works for diverse users
How Patterns Earn Their Spot
I don't document theoretical patterns. I document what's already working in products serving millions of users. This is design pattern mining: observing solutions in the wild and making them actionable.
3+ implementations
Works in multiple real products, not just one team's experiment
Real AI/UX problem
Addresses a fundamental challenge unique to AI-powered experiences
Actionable guidance
Every pattern includes code examples, demos, and implementation details
Research-grounded
Built on Google PAIR, Apple ML Guidelines, HCI research, and community practice
This is a living collection. Patterns evolve as products improve, new approaches emerge, and the community contributes insights.
By the Numbers
AI products analyzed
Validated patterns
Real-world examples
Strategic categories
Beyond Patterns
I've built additional tools to help designers apply these patterns in their daily work:
Gist.design
An AI design thinking partner. Clarify briefs, map user journeys, critique decisions, and prepare for stakeholder reviews all powered by the 36 patterns documented here.
Visit Gist.designFigma Make Prompts
36 copy-paste prompts for generating AI pattern components directly in your design files.
ExploreDesigner Guides
Step-by-step learning paths for AI design tools like Claude Code, Cursor, and GitHub Copilot.
ExploreOpen Source
This project is open source. If you want to suggest patterns, improve existing content, or contribute examples, I'd welcome it.
Get my weekly AI/UX analysis
Every week I break down one AI product's design decisions: what patterns they're using, what's working, and what I'd do differently. It's the analysis I wish existed when I started designing AI products.