My story

I started work on the factory floor, not in the office.

My first jobs were scheduling production lines, managing suppliers, and running warehouse logistics. I planned 4.5 million cans of infant formula a month at Danone. I scaled manufacturing throughput sixfold at EROAD to support their US launch. I led an engineering team at GEA. Before I ever ran a technology function, I spent years inside the operations that technology is supposed to improve.

That turned out to matter more than any qualification I've picked up since.

Most AI projects fail the same way. Someone builds a good model, wraps it in a clean interface, and deploys it into a workflow they've never actually done. Nobody in the room knows what a wealth adviser does between 9am and 11am on a Tuesday. Or which fields in the CRM the sales team actually trusts. Or why the compliance officer still keeps a separate spreadsheet. That gap between how the work looks on a process diagram and how it actually runs is where pilots go to die.

I've spent a decade closing that gap from the inside. Not as a strategist reading about these systems. As someone who has implemented them, sat with the teams who use them, and helped take these systems live.

I've led Dynamics 365 implementations across finance, HR, and operations at Datacom. I built a Salesforce managed services practice from nothing at Be Intelligent (now part of PwC), grew it to a quarter of their revenue, and held a 4.8/5 client satisfaction score across the book. At Fresh Direct I rescoped a $1.4M ERP down to $950K and redesigned the warehouse workflows that were costing the business 20+ hours of dispatch time a day.

Previously, I lead technology at Helm Wealth, a regulated wealth manager with about $5bn under management. I built the firm's first AI programme which shipped a production CRM agent on an MCP server inside Microsoft Teams. The agent connects to Dynamics 365 and helps the sales team prep for meetings, pull weekly summaries, capture leads, and add tasks. Everything went live with a privacy assessment and human-in-the-loop controls already signed off.

On top of the AI work, I rebuilt the data platform on Snowflake and Power BI (cutting frontline workload by 25%), set the technology strategy that earned a tenfold increase in investment from the board, led five acquisition integrations, rolled out a full cybersecurity programme, and delivered 5+ major projects a year on time and under budget with a small team and a $1M+ portfolio.

I also build on my own time. Not to learn AI. To test how agentic patterns hold up against real problems. DRONE: BREACH is a browser-based hacking game where a reactive AI antagonist adapts to your moves. Pit Lane to Production is a Fantasy F1 league on Substack where frontier AI models compete as team principals. I built a two-tier model-routing harness that keeps agent costs practical by sending only the hard queries to the expensive model.

I was named on the NZ CIO50 2023. I was a TUANZ Outstanding Future Leader finalist in 2024. I sit on TUANZ's AI Governance Committee and Unitec's Computer Science Industry Advisory Committee. I mentor through Summer of Tech and The King's Trust. I'm finishing the Oxford Executive Diploma in AI for Business this year.

If you're trying to get AI past the pilot, lets chat.