James Kitney

The AI spend conversation in NZ is mostly hypothetical. Most organisations haven't spent enough to have a real cost problem - they've read the horror stories and gotten scared. That fear is showing up as conservatism: restricted access, slow rollouts, tools sitting unused because someone's worried about a bill that hasn't arrived yet. AI inference costs dropped 67% last year. They'll keep falling. The pricing model is the short-term problem. The capability gap is the long-term one. NZ's had a productivity problem for decades and it's getting worse. Yet we have a tool in our hands that could actually move the needle - if we give people access to it. Your most expensive AI tool might be the one you were too scared to switch on.

Mark Zuckerberg has publicly admitted which parts of his CEO role he doesn't actually want to do anymore. He's building an AI clone to handle his internal comms so he can spend ten hours a week writing code instead. Most leaders won't be that honest, even with themselves. What I see across exec teams is the same pattern. We've spent two years pointing AI at the organisation. We've never pointed it at ourselves. The real cost isn't hours. It's the slow erosion of doing the work you love. A leader running on empty quietly caps engagement across the entire business. Zuckerberg's move isn't really about AI. It's about an exec willing to publicly name the parts of his role he no longer adds value to.

Most executive teams have the same routine. The monthly report arrives, the dashboard gets reviewed, the numbers get discussed. If someone wants to go deeper, a request goes to the data team, and the answer comes back days later — slightly out of date, and usually a compromise on the original question. So leaders stop asking. The curiosity doesn't disappear. It just goes unanswered. What AI makes possible now is something structurally different: the ability to ask a specific, pointed question of your business data in plain English and receive an answer in real time. No intermediary, no delay, no reformatting the question to fit what the system can already handle. The organisations building this habit are finding intelligence they didn't know they had — not because the data wasn't there, but because it was always too hard to get at. The next threshold for AI adoption isn't access. It's inquisitiveness. And the leaders who learn to ask better questions will surface insights that no quarterly report cycle ever could.

In 2026, leadership teams are facing a harder economic reality than expected, with stubborn cost pressures, uncertain demand, and growing limits to the traditional response of cutting costs harder. In that environment, the organisations that gain advantage will be those that use AI not as a standalone technology play, but as a strategic lever to redesign how work happens, offset external pressure, and deliver measurable commercial value.

Too many AI pilots in New Zealand are stuck in pilot more, and it’s time we addressed the uncomfortable truth, experimenting with tools is not a strategy. While the initial "buzz" of AI is exciting, the leap from a successful PoC to a truly transformative, scalable solution is where many of our organisations are stalling. This isn't just a technical hurdle, its a leadership and mindset gap that keeps us stuck in a cycle of experiments. In my latest article, I break down the five fundamental ideas required to move past the "shiny object" and build a culture of scalable AI… from planning for 100x adoption on day one to rethinking ROI beyond the pilot.

In this article, James Kitney explores how AI is eroding entry-level roles and reshaping traditional career pathways. He considers the rise of AI-enabled solo founders and new co-operative models as alternative responses. Ultimately, he challenges leaders to redesign strategy and organisational models for a radically different competitive future.