How I build with AI
I use AI to move between customer research, product design and software development. You work directly with me on direction, tradeoffs and delivery, with working examples to review and clear decisions about what is ready for users.
I work with the people who understand the customers, business and technology. We agree what needs to change, look at what’s already in place and decide what’s worth building.
Sometimes that means a new product. Sometimes it means improving an existing experience or workflow. The business and customer need determine the approach, including where AI is useful.
I turn the direction into designs, prototypes and working software you can respond to. Seeing the experience exposes questions a brief can miss and gives customers something real to try.
AI helps me explore alternatives and work through implementation. I review the output, make the tradeoffs and connect what we learn back to the goal.
Public examples from October 2026. The consumer AI product was in live alpha; the AI SaaS product was in progress.
On the consumer product, my AI testing agent worked through a user-test recording with the product and design agents. I had established direction and reviewed their work, then delegated the fixes they could make and the decisions that needed me. Six fixes were live that afternoon.
The useful result was a shorter path from observed friction to an improved experience.
A design practice from the consumer product prompted a large-text review on the SaaS product. It found six usability issues, corrected in the mockups that day.
A useful lesson became a concrete design improvement while each product kept its own context.
At Ticketmaster, I led global mobile product and worked directly in the codebases. I built prototypes for teams to evaluate and shipped production changes through senior engineering review.
AI search and interface prototypes gave teams and executives a concrete experience to evaluate before committing further. These remained prototypes.
Analytics and performance contributions shipped into the iOS and Android apps after senior engineering review.
For one home-screen module, I reduced three calls to one by linking to the artist page, where fans could choose an event. The change passed senior engineering review and shipped, bringing the module’s load time from about a second to 100 milliseconds.
AI lets me do more of the work directly. Expertise and clear ownership still matter. I review the work and involve the agreed technical owners in testing, integrations, data handling and release decisions.
A prototype, a user test and production readiness are different milestones. We agree which one we are working toward and what evidence will show it is ready.
We keep decisions, code and setup records current, and agree ownership, access, operating costs and ongoing support. You can see what has changed, what remains uncertain and what needs attention next.
I keep improving how I work. Useful lessons become part of my playbook, and I test them against the next problem rather than assuming they apply everywhere. Each project keeps its own data and permissions.
This October 2026 diagram shows how I organize AI agents across the builds: separate responsibilities, written handoffs, independent review and defined permissions. It describes AI tooling, not a staffed delivery team or a support commitment.
Download the working-setup overviewTell me what you want to change, who it would help and where you need a partner. We can talk through whether you need a software build, help alongside your team or advice on where to go next. A technical brief isn’t required.
Know a founder or team facing an important product decision or software initiative? An introduction is welcome.