Digital products and automation
I turn a problem into written rules and a plan for every screen, direct AI coding agents to build it, and review each build before it ships.
- Here
- Three projects: one live, one client prototype, one working demo on sample data
- My part
- Rules, design direction, the words, review of every build
- The code
- Written by AI coding agents I direct
The projects, and my part in each
-
Public websiteLive
Brampton Civic Signal
A live, independent guide to Brampton’s 2026 election, built on one rule: every fact names its official source and the date it was checked.
- My part
- Wrote the publishing rules, directed the design, approved every release
- Made with
- AI coding agents I directed, in Astro and TypeScript
- Stage
- Live at bramptoncivic.ca
-
Customer journeyClient prototype
Another Level
Twelve screens a family meets with a Brampton youth sports business, rebuilt around the owner’s written priorities.
- My part
- Turned the owner’s brief into the journey, directed every screen
- Made with
- AI coding agents I directed, in Next.js, React and TypeScript
- Stage
- Interactive prototype, shown with the client’s permission
-
AutomationDemo, sample data
Lead Recovery Command Center
A working demo that catches a business’s missed calls, texts and forms, replies to routine ones, and waits for the owner when it cannot reply safely.
- My part
- Designed the stages and decision rules, and shaped the demo
- Made with
- AI coding agents I directed, in Next.js, React and TypeScript
- Stage
- Working demo on sample data
From a problem to something you can use
-
Start from the problem as the client writes it
For Another Level, the owner set the direction in writing: two divisions, reusable athlete profiles and an order for the work. I rebuilt the prototype against it.
See the owner’s brief -
Write the rules before the screens
Brampton Civic Signal runs on one rule: every fact names its official source and the date it was checked. If a fact cannot be sourced, it does not appear.
See a source receipt -
Direct AI coding agents to build it
Agents write the code, in Astro, Next.js, React and TypeScript. I set the structure, the design direction and the words on every screen.
-
Review every build, and rule on what is wrong
I review each build at phone and desktop size. In July 2026 I changed the demo’s main rule: routine replies go out on their own, and anything risky comes back to the owner.
Follow two leads through it -
Ship something a person can use, and say its stage
Of the three here: one live, one client prototype, one working demo on sample data. Each page says which.
Hiring for a role? My experience and tools are on About. For a website, an app or an automation for your business, get in touch.