AI The Right Way™: starting from an informed, educated foundation

Most organisations come to AI the wrong way round. They start with a tool, a licence, a demo, a free trial, and only later, if at all, ask what they are actually trying to achieve, whether their people understand it, and whether it is being used safely. AI The Right Way™ starts from the opposite end: from an informed, educated foundation. Understand the technology first. Then decide where it belongs, govern how it is used, and apply it with intent.

AI The Right Way™ is the trademarked methodology we teach, developed with The AI Assembly out of years of experience watching what organisations and their people actually need to adopt AI safely and well. It is the foundation of every course we run, courses that have already taught these fundamentals to more than 1,500 people across New Zealand and Australia. It is not about fear, and it is not simply about cost. It is about confidence: people who understand what they are using, and therefore use it well.

What “the right way” actually means

Doing AI the right way is not about fear or red tape. It means four things working together. It is informed, built on a genuine understanding of how the technology works and where it fits. It is responsible, used with honesty about its limits and real care for your data and your people. It is governed, with clear policy, ownership and guardrails. And it is purposeful, used for defined tasks, not an unmonitored free-for-all. Start from an educated footing and these become natural; skip it, and they all become a struggle.

Why an informed start matters more than ever

Starting from an educated footing has always been the sensible path. What has changed is the cost of not doing so. The ground beneath these tools is shifting fast, and an uninformed start now carries real consequences, in wasted effort, exposed data, poor decisions, and, increasingly, money.

Consider what happened on 16 June 2026, when Microsoft made Copilot Cowork, its agentic assistant that carries out tasks such as drafting documents, building spreadsheets and sending emails, generally available, and switched on usage-based billing. On top of the existing Microsoft 365 Copilot licence (around US$30 per user per month), every task now consumes Copilot Credits, priced on four moving inputs: the model used, the context retrieved, the tool calls made, and the runtime. Teams that had been exploring these capabilities through Microsoft’s Frontier programme since late March were moved onto consumption billing, in what Microsoft has described as its first pricing-model change in nearly two decades.

The point is not the price. It is that the rules can change overnight, and an organisation working from an informed foundation can absorb a shift like this calmly, while one that started with the tool, and never built the understanding, cannot. This is the first of several moves like it, and we will be writing more on the cost of use and what it is likely to look like as the market settles.

What this looks like in everyday use

Consider one ordinary habit: iterating with an AI for an hour, nudging it toward the right answer one prompt at a time. Under a flat subscription, that hour was simply time. Under pay-per-action pricing, that same hour is now a cost, every attempt, every regenerate, every half-formed prompt adds to the bill. Multiply it across a team that was never taught to plan a prompt, and the numbers climb fast.

This is exactly where the fundamentals we teach come home to roost. Knowing how to plan a task before you start, instead of iterating blindly for an answer. Knowing how to align AI to a defined workflow, rather than reaching for it on everything and anything. Knowing how to audit what is being used, by whom, for what, and at what cost. These were always good practice. With per-action pricing and no clear cost mechanism, they have become the line between AI that pays for itself and AI that quietly drains the budget.

Why controls alone will not save you

The instinct, when costs and risks rise, is to lock everything down, restrict access, add approval gates, block tools. But over-restrict and people either cannot do their jobs or quietly route around you with personal accounts, which is far worse. You cannot write a policy for every situation an employee will face. What actually protects you is judgement: people who understand the tools well enough to make good decisions in the moment. Judgement is not installed. It is taught.

How we teach it, and embed it

AI The Right Way™ is the framework. Educate to Enable is how we put it to work inside your organisation. Working alongside private businesses, we follow a simple sequence: educate, enable, embed. First we build genuine understanding and confidence. Then we enable your people to apply it safely to real work. Then we embed it into your ways of working, your processes and your wider organisation, so it sticks long after we leave. Understanding comes before strategy. Strategy comes before spend.

We deliver the methodology through focused half-day and full-day courses, each aimed at a different part of your organisation:

AI Foundations (half day, all staff), what AI actually is, what it is not, how it is priced, and where it creates real value. This is where everyone gains the shared language and basic cost awareness that prevent expensive mistakes.

AI for Leaders (half day, leadership), how to govern AI adoption, set policy, manage risk, and make smart investment and budget decisions. Leaders leave able to put the right guardrails in place.

Prompt Engineering Masterclass (full day, knowledge workers), how to get dramatically more value from the tools you already pay for. Skilled users get better results with fewer wasted, costly attempts.

AI Maturity Assessment (two hours, leadership and operations), an honest read on where your organisation sits today, and where the real risks and opportunities are.

Industry Deep Dive (full day, sector teams), the responsible, high-value use cases specific to your sector, and the pitfalls to avoid.

Aligning AI to how your organisation actually works

Education is the foundation, but on its own it is not enough. AI done the right way also has to fit the organisation it lands in. This is why a clear roadmap matters so much: it aligns where and how AI is used with your actual structure, your ways of working and your people, rather than bolting a tool onto processes that were never designed for it.

Done well, this is where the real gains show up. Auditability is built in rather than bolted on. Inefficient ways of working are redesigned around what AI is genuinely good at, instead of automating a broken process. Governance lines up with how decisions are actually made. The result is AI that strengthens the organisation, because it was planned around the organisation, and not the other way round.

Everyone has a part to play

A rogue user is rarely malicious. Usually they are simply unaware, pasting sensitive data into the wrong tool, running an expensive process on repeat, or trusting an output they should have checked. The fix is shared responsibility, and every level owns part of it.

Leaders own governance: the policy, the budget, the guardrails, and the decision about which tools are sanctioned. Team leads own oversight: making sure their people use approved tools for defined tasks, and flagging issues early. Employees own everyday judgement: using AI for what it is meant for, knowing roughly what it costs, protecting data, and checking the work. When all three are educated and aligned, the rogue-user problem largely disappears, because far fewer people are operating in the dark.

Govern, audit, and control the costs

With everyone on board, the practical controls finally work. Governance becomes a living policy people understand and follow, not a document nobody reads. Auditing becomes possible, because you know who is using what, for which tasks, and at what cost, so spikes get caught early. And cost control becomes a shared habit rather than a finance-team firefight: people who understand usage-based pricing naturally use these tools more deliberately.

The cheapest insurance you can buy

When one careless session can cost thousands, the maths on education is simple. Training your people is not an expense set against AI, it is the thing that makes AI safe to spend on at all. The organisations that will thrive as costs shift are the ones that got everyone on board early: leaders, managers and employees alike, all speaking the same language and pulling in the same direction.

That is what AI The Right Way™ delivers, taught and embedded through Educate to Enable, and it is where every engagement we run begins. If you want your whole organisation ready for what is coming, before a surprise bill makes the case for you, get in touch and we will help you start.