There is no simple guide to AI transformation, just adaptable leaders
Fri, 11th Sep 2026 (Today)
There is a falsehood in the way we sometimes talk about AI transformation: that there is a simple guide leaders can follow.
There isn't.
Not when the technology keeps getting faster. Not when the best use cases are still being discovered. Not when yesterday's assumptions can be outdated by the next model release.
The hard part of AI is no longer the technology. It is changing how work gets done while earning and maintaining people's trust - with customers, employees, regulators and the communities organisations serve.
Our recent AI Trust Report shows the tension clearly: New Zealanders are increasingly using AI, but they are also asking harder questions about whether it is being used responsibly, transparently and with the right human oversight. That makes trust a leadership issue, not just a technology or compliance issue.
The foundations underneath AI are too often underestimated.
As soon as you want AI to do something genuinely useful inside an organisation, it needs first to understand your business, customers, people, products, rules, language, what good looks like to you as an organisation and what it must never do. This is why the real work often sits in the less glamorous foundations: process, knowledge, data, memory and context. Here are three lessons we've learned at One NZ as we transform our business for an AI-powered future.
- Hire a Transformation Leader but never outsource a transformation.
Partners are incredibly useful. They bring expertise, pattern recognition, capacity and speed. You can outsource technology, but you cannot outsource the mindset, ownership and will, required to become an AI-enabled organisation.
The context is yours. The judgement is yours. The customer is yours. The operating model that needs to change is yours.
Ultimately, the people who need to own the agents, workflows and outcomes are the people running the business. They are the ones who can create, maintain and improve the transformation over time.
Bring a Transformation leader in by all means but don't outsource your way through PoCs and projects. The strategy, the change and much of the capability build needs to be your own. This is how you end up with a really transformed company, not just a collection of impressive AI projects.
- Learning is the playbook. Safe experimentation is the strategy.
When One NZ decided to start on the transformation journey, there was no playbook or a guide to follow. We needed to learn how to build, experiment to understand what was working, and operate long enough to see what we needed to unlearn.
We have repeatedly taken a step forward with AI, then a step - or more - backwards to sort out the knowledge, data or process underneath it. AI at an enterprise strategy level can feel exciting, but how it can be integrated in roles and tasks, that can feel very different.
Getting ahead in AI has as much to do with knowing your organisation as it does with knowing the technology – if that's unclear your AI efforts will stall.
Leaders must get closer to the work, going down into details to the level of tasks and workflows sometimes to find exactly how the work is being done. The closer you get to someone's actual work, the more real AI becomes.
- Focus on adoption and building capability.
Give everyone access and make it safe to experiment, then put disproportionate effort behind the people who are curious, impatient and want to run with it. They create examples for the rest to see and come along. That's the real transformation.
There is always a risk that capability develops unevenly. The answer is not to hold everyone back, but to create safe access for all while backing the early movers who can show what is possible.
Instead of automatically putting a large specialist group around every problem, we can ask a more useful question: what could a smaller, well-supported group of capable people do if each of them was using AI well?
They can adapt and shift their mindsets, moving between functions and ways of thinking, to fill the gaps with AI. That's a continuous process.
My final advice: just get going.
The next phase of AI leadership will not belong only to the technologists. It will belong to leaders who can stay curious, ask better questions, redesign work responsibly and build trust as they go.
There is no simple guide. That may be the point. The organisations that succeed will not be the ones waiting for the playbook. They will be the ones learning fast, moving carefully where trust matters most, and treating trust as part of the transformation - not something to bolt on at the end.