I am Robert Marsch, a product designer. I design AI-native products and run the delivery that ships them as working code, not just prototypes.
UX research and AI-native workflows, turned into products that ship as production code.
I design products for complex, multi-stakeholder problems.
I'm a product designer and product owner working on complex, multi-stakeholder digital transformations for government and enterprise clients. I pair rigorous UX research with AI-enhanced workflows to ship scalable design systems, product roadmaps, and the delivery that puts them into production.
Recently I led product strategy for government tourism platforms serving 800,000+ monthly visitors, ran pod-based AI delivery for a Middle East resident super-app, and managed cross-functional teams of 20+. I care about the boundary where design becomes working software.
- AI-native product design and delivery
- UX research and CX strategy
- Design systems, shipped as code
- Multi-stakeholder digital transformation
Five products, one ecosystem.
The client is a leading Middle East real estate developer. We rebuilt the resident lifestyle app its communities use every day: property requests, payments, education fees, rewards, and more. Five fragmented products merged into one AI-native ecosystem, 400+ screens shipped as production-ready code in six weeks.
Five fragmented products, each with its own patterns, and residents felt the seams.
The client's challenge: rebuild the whole thing as one ecosystem, AI-native, in six weeks. No traditional agency process could hit that.
The team was restructured around the AI workflow.
Eight designers, new roles: a Delivery Lead on rhythm and quality, an Experience Architect on vision, an AI Operator on the system, designers on the screens.
Weeks 1 to 3: align the design system with reality. The live app and Figma converge, and the agent closes each gap with a designer in the loop.
Ideate, Pull, Verify, Build, QA, Ship. The prototype is the brief. One ticket, one session, one PR.
The hard part wasn't the tech. It was not going back to the old way.
Midway through, the client's internal teams pushed to fall back to manual Figma production, the way redesigns had always been done. As delivery lead I kept the pod on the AI workflow. Reverting would have lost the compounding gains that made the six-week timeline work.
The tooling is repeatable. Staying with it under pressure is what got the work shipped.
Every screen rebuilt. Delivered as code, not prototypes.
React Native styled with NativeWind. Files the client's engineers could open, review, and take straight into production.
“This is real transformation. This team is in the top 10% of AI adoption globally.”
CLIENT HEAD OF DIGITAL, NAMING THE POD AI ADOPTION CHAMPION AHEAD OF THEIR OWN TEAMS
Other case studies
Earlier research and strategy work across gaming, femtech, consulting, and construction.
UX audit for a major game studio
An expert usability audit of a legendary gaming company's website, run with no direct user access. Nielsen heuristic scoring, competitive benchmarking across six brands, and content analysis, distilled into a ranked issue list and concrete fixes.
An AI femtech platform for menopause
An AI-powered platform for menopause, a phase that affects nearly half the population but still lacks real support. A one-month sprint: user and provider interviews, a 200-story backlog, five personas, and 80+ proof-of-concept screens, built for empathy and whitelabel flexibility.
A digital workshop platform for a consultancy
A Singapore consultancy's in-person maturity workshops, taken fully digital. Startups answer maturity questions, the platform scores them, and consultants get quantifiable evidence. Designed for three audiences at once: founders, VCs, and facilitators.
From support tickets to self-service
A UAE villa construction platform where 40% of users filed a support ticket just to keep going. An AI-run pipeline produced a 57-issue heuristic audit across 17 flows, detailed PRDs, and 165+ redesigned screens, turning support tickets into self-service.