Consulting — AI-Directed Development
Most businesses can see what AI can do. The hard part is knowing how to direct it well enough to ship something real. That's the gap I fill.
The Offer
I'm a non-technical founder. I've never written production code. I have built a live SaaS product, a shipped game narrative portfolio, and a multi-tool desktop app. All of it by directing Claude, Codex, and custom AI agent pipelines as the engineering lead.
The skill isn't prompting. It's product thinking: knowing what to build, how to decompose the work, what to check at every step, and when to push back on what the AI produces. It's the same discipline that makes a good engineering manager — except the engineers are AI systems.
That process is repeatable, teachable, and immediately applicable. If your business wants to ship something with AI but doesn't know how to direct the work, that's exactly what I do.
Proof
An automated website builder for NZ tradespeople. A client describes their business, and SiteSorted generates and deploys a production website. Fully automated from form submission to live site. Built on Next.js 16, React 19, Tailwind 4, and the Anthropic Claude API. Cost per site: $3–7 in API credits.
sitesorted.co.nz — LiveA complete game narrative portfolio built in 8 weeks under AAA industry mentorship. 150+ character barks, 38 dialogue nodes, 6 unique quest endings, full world-building documentation, storyboards, and VO scripts. Every document designed for real team handoff.
Narrative portfolio — CompleteA Windows productivity suite: FocusDim, QuickBoard, LiquidFocus, SoundSplit, ScreenSlap, MeetReady, and a Dashboard launcher. Electron main process, Rust native sidecar, typed IPC bridge. A multi-process desktop application with zero lines of code written by hand.
Electron + Rust — In developmentBackground
Ten-plus years in enterprise roles at companies including Miro — leading cross-functional teams, managing complex stakeholder relationships, and delivering under tight deadlines. Former Account Executive with a strong product sense and a background in technical sales.
That foundation matters here. Directing AI to build software isn't just about knowing the tools — it's about knowing what good looks like, holding quality gates, catching failure modes early, and translating business requirements into something a system can act on.
The enterprise years built the product sense. The AI years turned it into a delivery capability.
Who It's For
AI tools have lowered the cost of building software dramatically. The bottleneck is no longer access to the tools — it's knowing how to direct them well enough to produce something that actually works in production.
If your team is stuck in the proof-of-concept loop, or you're watching competitors ship things you know you could build, this is where I come in.
Engage
No intake form, no discovery call template. Send me a message that describes what you're trying to ship and where you're stuck. I'll tell you whether I can help and what that looks like.