(Case Study)
AI-Native Product Design Lab
How I use AI to accelerate research, design, prototype, and delivery — built on years of bringing teams together.
- Personal lab + selective client integration
- 2022–present
- AI prototyping · Methodology · Custom GPTs · Vendor-agnostic advisory
- Claude · ChatGPT · Midjourney · Runway · v0 · Base44 · React · Next.js · Tailwind
The challenge
Most consultants advising on AI today are vendor-coded or theoretical. Few have actually built AI-augmented products end-to-end. Enterprise teams hiring AI advisors are getting slide decks and vendor partnerships. They're not getting people who can walk into a room, assess where AI actually fits, and demonstrate what's possible by building it in real time.
I built this lab to keep AI fluency hands-on — not because I needed to add AI to my marketing, but because I wanted to know, from the inside, what the tools could actually do in a regulated, healthcare-adjacent context.
The approach
I'm classically trained — Pratt BFA in Communications Design and Advertising/Marketing, fine arts background, two decades as a graphic designer, marketer, and senior UX leader. AI experimentation began in 2022, alongside a creative practice that dates back to college. My best friend, now Poet Laureate of Connecticut, and I co-founded a creative arts and music collective at Rider University. We've been collaborating creatively for 25+ years, getting together annually for live performances where I create live visual art alongside his poetry and music.
In 2022 I started using Midjourney and Runway to generate live visuals during those performances — projected through two or three projectors, tied to lyrics and music in real time. Live performance is an unforgiving classroom for AI tooling. There's no regenerating when the band is playing. Everything I now know about how AI fits into design work started there.
The lab has expanded since then into healthcare-specific prototypes and tools.
AI Patient Support
The future of healthcare is disease-first, brand-second.
The AI Patient Support concept is the working prototype that proves it. Instead of patients searching across twelve brand sites for assistance, they search by their disease and find every option in one place — every support program, savings card, study, KOL video, and patient story, verified and organized by condition.
The concept covers Oncology (4 indications, 47 drugs), Diabetes, Cardiovascular, Immunology, Neurology, HIV/AIDS, Rare Disease, and Respiratory. Click a condition, find every drug. Click a drug, find every resource — support and savings, videos and media, downloadable materials, clinical studies, drug timeline, legal and safety. Each piece of media tagged by length, source, and category.
Channel Optimizer
An AI-assisted dashboard for media-mix decision support, built in v0. Channel performance, engagement metrics, and AI-generated insights and recommendations — “Increase budget allocation to social media campaigns by 15%,” “Consider reducing radio ad spend and reallocating to higher-performing channels.”
Built in 4 hours. Would have taken a 4-week designer-engineer sprint in 2022.
HealthSync Pro
A health-data integration prototype built on Base44 — Garmin watch device configuration with regulatory-grade granular consent design. Heart rate, steps, sleep tracking, workout data, blood pressure, blood oxygen, body temperature, weight metrics — each with explicit user permission toggling.
This is what AI-assisted prototyping looks like for healthcare specifically. Consent design isn't a checkbox feature. It's the product.
Apple Vision Pro Concept
In-office HCP education rendered inside a Vision Pro spatial computing interface. Patient Weight Over 5 Years chart, patient profile, SD-tier navigation tabs, and a Type O- Universal Donor blood-type badge — designed for the moment when an HCP and patient are reviewing data together in an exam room.
Most senior pharma designers haven't touched Vision Pro yet. The frontier isn't where you wait to arrive. It's where you go.
Three Custom GPTs
Three working Custom GPTs deployed publicly:
- UX Research Advisor — advises on UX research and marketing opportunities using user-provided data.
- Product Innovation and Case Study Helper — generates detailed product definitions and case studies.
- AEM Design Assistant — best practices for AEM, Adobe Target, and DAM for storytelling and personalization.
Each is live in ChatGPT. Anyone can click through and use them right now.
Automation Opportunity Assessment
Built for teams asking the real question — where does AI actually fit in our workflows, and how do we know it'll pay off?
The Automation Opportunity Assessment is a six-part audit framework that moves an organization from “we should be doing something with AI” to a prioritized, feasibility-tested roadmap. The structure: Workflow Audit (where the friction actually lives) → Feasibility(what's buildable now versus next year) → Impact (where the ROI lands) → Agent Logic (the decisions the system needs to make) → Zapier Stack (the integration spine) → Reflection(what we learned, what we'd do differently).
This is the deliverable I run inside Advisory engagements when a leadership team is being pitched AI by every vendor in their inbox and needs a vendor-agnostic read on what's real for their workflows. The framework was developed during a Zapier AI Automation course and refined through hands-on engagements. It's the bridge between the individual prototypes above and a systematic AI strategy for an enterprise team.
Education Tools for Daughters
The lab also includes a small repository of educational apps I've built for my daughters — math and English learning tools that started as personal projects and have grown into a free at-home learning supplement. Same instinct as everything else: the right tool, in the right moment, for the actual person who needs it.
“Now any idea you can conceive, you can create stimulus to test. That speed-to-market is real. AI has become the everyman's opportunity to design and create — and the people who'll win are those with subtle nuance, who can tell good from better from great. My AI Native Lab work is about three things: efficiencies in my own work, the ability to help people, and the ability for the companies I work with to generate revenue.”
Selected outcomes
Working prototypes built in hours, not weeks
AI Patient Support, Channel Optimizer, HealthSync Pro, Apple Vision Pro concept — all built using AI-assisted development tools (v0, Base44, Claude, ChatGPT).
Three Custom GPTs deployed publicly
Working AI tools live in the world, not concepts in a deck.
A vendor-agnostic AI advisory practice
Built on hands-on tool experience, not vendor partnership commissions.
25+ years of creative practice extended with AI
Not replaced by it. Annual live performance work continues, with another AI-assisted performance planned for August 2026.
A repeatable methodology for integrating AI into team workflows
The 10 Source Packs framework feeds AI the inputs it needs to produce real strategic work, not generic output.
What this means
Most teams hiring AI consultants get advice. The teams hiring me get advice plus a demonstration.
AI isn't a strategy. It's a tool. The teams that win with AI long-term aren't the ones with the best models. They're the ones who treated the experience architecture around the model as the actual work.
That's the lab. That's what I bring into client engagements. That's the difference.
Related work
Pfizer · Multi-Brand · Regulated Industry
Building digital governance across 15+ therapeutic brands
$3.5M+ in digital transformation. Industry-first mobile wallet integration for patient medication information.
Chrysalis Initiative · Patient Experience
A patient navigation platform for women facing bias in cancer care
D&AD Pencil 2022 for Future Impact. Two-sided platform connecting patients with trained coaches and peer navigators.