Almost every business in New Zealand has now switched on artificial intelligence in one form or another. Far fewer have learned to use it well, and the gap between those two things is quietly becoming the most important competitive line in the country.
The headline numbers look encouraging. A study prepared by Deloitte Access Economics for 2degrees found that small and medium businesses using AI earned roughly $400,000 more in the 2025 financial year than comparable firms that did not, while large adopters were ahead by about $59.1 million. Eighty-two per cent of the businesses surveyed said they use AI in some way. The Government has built a whole national strategy around that promise, releasing its AI strategy, Investing with Confidence, in July 2025 with a deliberately light-touch approach and pointing to research that suggests AI could add around $76 billion to GDP by 2038.
So far, so optimistic. Look a little closer, though, and the picture gets considerably more honest.
That 82 per cent figure hides how shallow most of the adoption really is. The Deloitte research notes that a large share of businesses are simply using AI features that came bundled inside software they already owned, rather than genuinely rebuilding how they work. Andrew Fairgray, chief business officer at 2degrees, put it plainly when he said the data shows AI is already in use every day across businesses. Using it and getting value from it are not the same thing. Liza van der Merwe, who led the work at Deloitte Access Economics, made the point that the real gains come from “building mindset, systems and skills in tandem”, which is a long way from switching on a feature and hoping.
Older official numbers tell the same story from a different angle. A Datacom survey cited in the Government’s own state of play analysis found that 67 per cent of larger New Zealand businesses now use some form of AI, up from 48 per cent the year before. The smaller end of town is a different world. An NZIER and Spark survey found 68 per cent of small and medium enterprises had no plans to even evaluate AI, let alone invest in it. For comparison, only 38 per cent of Australian small businesses said the same. We are not short of enthusiasm at the top of the market. We are short of depth, and short of the skills to build it.
The part that should concern any business owner most, though, is what their customers think. One NZ released its AI Trust Report 2026 this year, and it reads like a warning shot. Around 76 per cent of New Zealanders used an AI tool of some kind in the past year, so exposure is no longer the issue. But 70 per cent of those users ran into problems with it, and 62 per cent said they would stop using a product or service over concerns about how a company was using AI. Sixty-eight per cent want a human option to fall back on when an AI interaction goes wrong. Summer Collins, One NZ’s chief AI and business services officer, summed up the mood neatly when she said “exposure doesn’t equal acceptance, if anything the opposite is true”. The more people use these tools, the more clearly they see the limits, and the less patient they become with businesses that hide behind them.
Put those three findings side by side and a clear pattern emerges. Adoption is high and rising. Genuine capability is thin. Customer trust is hardening rather than softening. That is not a comfortable combination, and it is exactly the situation in which a lot of money gets spent badly.
From where I sit, running a web and digital agency, the businesses that are actually pulling ahead are not the ones that bought the most AI. They are the ones that picked one real workflow and rebuilt it properly. A builder who used to lose two evenings a week writing quotes, now drafting them in twenty minutes and checking them carefully before they go out. A retailer whose product descriptions used to sit half-finished for months, now written in a morning. These are not glamorous transformations, but they compound, and they are the kind of thing the Deloitte revenue gap is really measuring.
The failures tend to share a shape too. A chatbot bolted onto a website with no human escape hatch, answering confidently and wrongly until a frustrated customer gives up and rings a competitor. Marketing copy generated in bulk, published without a human read, and quietly eroding the trust of the exact audience it was meant to win. The tool is rarely the problem. The absence of judgement around it almost always is.
If you take one thing from the research, make it this. Treat your website and your customer-facing systems as the place where AI either earns trust or destroys it. Be transparent when a customer is dealing with a machine. Always leave a clear path to a real person, because two thirds of your customers are now explicitly asking for one. Keep a human reading anything that carries your name before it reaches the public. And resist the temptation to measure success by how much AI you have switched on, rather than by what it actually does for the people you serve.
New Zealand has, on paper, leapt onto AI faster than almost anyone expected. The next two years will be decided by something less flashy and far harder to fake, which is whether we bother to learn how to use it. The businesses that close the gap between switching it on and using it well will quietly take market share from those that never do.
Roger McSaveney is General Manager of Forge Creative, an Auckland web design and development agency.
What is your experience with AI in your own business or workplace? Has it genuinely helped, or just added noise? Share your thoughts in the comments below.
This article was written by AI, briefed to report the facts, hopefully without some of the bias people bring to the job 🙂
Leave a Reply