Joel Lipton
AI Systems &
Product Design
Product designer focused on AI systems and how people work with them, make decisions, and stay in control.
How I work with AI
How I use AI to prototype quickly, carry context between projects, and build checks into my workflow.
Featured Work
01–03
Healthcare / AI interaction design
Designing an AI Copilot for Hospital Safety Investigations
Designing personalized AI assistance that helps hospital safety teams analyze incidents, inspect evidence, and stay in control.
Secured client buy-in by setting the vision for AI-powered incident management
View case study
Healthcare / Foxcommand.ai
Finding the Right Place for AI in Hospital Safety
I worked directly with patient safety directors at a Regional Health System to understand core workflows, identify the highest-leverage starting point for AI, and turn that strategy into a working product.
Amazon / AI Shopping
Designing an AI Agent for High-Stakes Commerce
I designed and shipped an AI shopping agent that helps customers navigate complex purchases and make better decisions before they buy.
More selected work
03–04
Amazon
Creating an industry first Multi-Box Order Experience
$100M reduced concessions
Designing systems that manage multi-package shipments across Amazon’s logistics network.
View case study
Amazon
Designing Service Discovery When Add-to-Cart Moves to Search
Increased service attach rate
I introduced service discovery and purchasing into search for the first time, increasing service attach rate and engagement.
View case studyApplied AI Product & Systems
Designing and building AI products, interaction models, and systems for real-world use.
I Built a Working Earthquake Explorer App in Under 48 Hours
What I learned taking Seismos from idea to working product in 48 hours. A look at how AI changes speed, where design judgment still matters, and why the last 20% is often the hardest part.
Read on Substack
Reconstructing AI Decisions in Healthcare
As AI systems begin influencing clinical coding and claim submissions, organizations must be able to reconstruct how those decisions were made months later. This prototype demonstrates execution-layer governance that preserves AI decision artifacts and enables deterministic replay during claim appeals and compliance reviews.
Read on Substack