We’ve been teaching the fundamentals for 24 months and it’s starting to pay off.
June 2026
Two years ago we made a choice that, at the time, looked unfashionable. While much of the market was selling tools, demos and the promise of overnight transformation, we did something quieter: we taught people the fundamentals. What a token actually is. How a model really works, and where it does not. Why context matters, what these systems can and cannot do, and how to tell genuine value from hype.
It was not the flashy option. But we believed then what we know now: you cannot get value from a technology your people do not understand.
And we did it at scale. Over the last two years we have taught these fundamentals to more than 1,500 people across New Zealand and Australia, leaders, teams and frontline staff alike. At the time, plenty of it looked like patient groundwork with no obvious return. That groundwork is exactly what is now starting to pay back.
Why we started with tokens
A token is the small unit of text an AI model reads and generates, roughly a few characters, or part of a word. It sounds like a technical detail. It is not. Once someone understands tokens, a lot of AI stops being magic and starts being something they can reason about: why a model has limits, why long documents need to be handled with care, why cost and speed move the way they do, and why a clear prompt beats a vague one.
Teaching tokens was never really about the word “token”. It was about giving people an accurate mental model, the foundation everything else is built on. Understanding before strategy. Strategy before spend.
What changed
Through 2025 and into 2026, AI stopped being something a handful of enthusiasts experimented with and became something organisations are expected to run. Agents that take actions, not just answer questions. Higher stakes, real money, real risk, real governance questions. And the gap between organisations that understand what they are deploying and those that do not has widened sharply. It is now visible in the results.
The teams who learned the fundamentals are the ones moving with confidence. They know which problems are a good fit for AI and which are not. They write better prompts, design better workflows, spot the failure modes before they cause harm, and ask vendors the right questions. The teams who skipped the fundamentals are still buying tools nobody uses, surprised by costs they did not predict, and nervous about risks they cannot quite name.
The fundamentals are the moat
Here is what the last two years have made obvious: tools change every few months, but understanding compounds. The organisation that grasps how these systems work can adopt any tool, adapt to any shift, and keep finding new value long after the current hype cycle has moved on. Capability is the one thing a competitor cannot buy off the shelf.
That is why we built Webbased AI the way we did, Learn, Plan, Build, in that order. Not because training is a nice add-on, but because it is the thing that makes everything after it actually work.
It is not too late to start
If your organisation skipped the groundwork, the good news is that the fundamentals have not changed, and they are not hard to teach well. A team can go from nervous to genuinely capable far faster than most people expect, and once they do, the tools, the roadmap and the results follow.
We spent two years making the unfashionable bet that the fundamentals would matter most. That bet has come in. If you want your people on the right side of the gap, that is exactly the work we do, get in touch and we will show you where to start.