Nic Fouhy
AI Integration Consultant
Kapiti Coast, New Zealand
Sixteen years scaling a national operation
repairs/year when I started
cities nationwide
staff at peak
repairs/year at scale
I spent sixteen years at New Zealand's largest privately owned technology company. When I joined it was a team of two doing around 200 repairs a year. By the time I left, we had 34 staff across four cities handling over 27,000 devices annually. I became a shareholder along the way. We built the largest technology repair facility in New Zealand.
During those years I built the software that made that scale possible. Custom ERP systems. API integrations across insurance and logistics platforms. A computer vision ML model trained on 216,000 damage images achieving 95%+ accuracy. These were not side projects. They were the tools that let a two-person operation grow into a national one.
Logic gates for ambiguity
I think like a programmer. Logic, states, functions. Layers of calculations solving problems, visions brought to life in code. The part that hooked me was always the same: clicking run and watching a collision of checks, inputs, maths, and live data from multiple systems resolve into an instantaneous result. That is magic to me. Magic I made.
When AI language models matured, most people were captivated by the writing. The creative output felt like the point.
My background made me see something different.
Words could now be used as logic. That was the real shift.
A traditional programme can check whether a number is greater than ten. It can match a string exactly. But it cannot tell you whether a customer is frustrated or simply asking a question. It cannot work out whether a maintenance request implies a health and safety risk that was never explicitly mentioned. Before AI reasoning, that kind of information was too messy to code against.
Now you can. AI reasoning lets you output true or false from language that is fuzzy, implied, and never directly stated. You can treat language nuances as variables. Convert unstructured text into structured decisions. Build functions that extract what a message meant to say, not just what it literally said.
That opened an entirely new class of programmatic functionality. Every project at EmbedAI starts from this premise: if you can describe the logic, AI can now execute it, even when the inputs are human language at its most ambiguous.
What I build
Every EmbedAI project is, at some level, about applying structured logic to information that used to resist structure.
By the numbers
years building operational systems
device repairs per year at peak
damage images trained on
Smart Assess accuracy
Built production systems for IAG, Vero, Tower, property management firms, HVAC companies, plumbing businesses, and joinery workshops across New Zealand.
How I work
I am not interested in AI for the sake of AI. I am interested in what it can do inside a real business with real constraints and real customers. If a project does not have a clear return, I will tell you that before we start.
The companies I work with are not startups chasing hype. They are established operators who want to automate the parts of their business that cost them time and money.
EmbedAI is based on the Kapiti Coast. I work directly with clients. No account managers, no layers between us. When you ring, you get the person who built your system.
Let's talk
If you have a process eating hours, a phone that goes unanswered, or a manual workflow you suspect could be automated, I am happy to talk it through. No pitch deck. Just a conversation about what is possible and what is not.