If 91% of firms use AI, why is real transformation falling?

The headline number is the wrong number

Every press release will lead with it. AI adoption has hit 91% of New Zealand businesses, according to Datacom’s fourth annual AI Index survey, published on 28 August 2026 and based on 207 senior managers polled in late July and early August. It sounds like a country sprinting into the future.

Then you read the number underneath it. Only 4% of businesses are using AI to transform core operations, down from 8% the year before. Adoption doubled its reach over three years while genuine transformation halved in one. That is not an innovation story. It is a productivity gap dressed up as one.

What most firms are actually doing

Datacom’s AI director Lou Compagnone has a name for it: “AI confetti”. Individual staff using the technology for ad hoc improvements in their day-to-day work, with company-wide projects still rare. The survey backs this up. Only 15% of organisations are scaling AI organisation-wide, and 68% are relying on general-purpose assistants such as ChatGPT, Gemini and Claude, by far the most common technology in use.

A July 2026 Newswire analysis explains why the headline take-up rate is so high and so misleading. When a business reports “using AI,” it usually means someone clicked a summarise button in their email or accepted a draft suggestion in a word processor. Useful, frictionless, and almost invisible on the bottom line. As the analysis puts it, “bolting an assistant onto an unchanged process makes that process slightly faster. It does not make the business meaningfully more productive.”

Deloitte Access Economics found the same pattern in its study for 2degrees, with 82% of businesses using AI in some form. Its chief business officer Andrew Fairgray put it bluntly: “Using it and getting value from it are not the same thing.”

The gap between dabblers and transformers is now measured in millions

This matters because the difference is not academic. Deloitte’s April 2026 report for 2degrees found average SME AI adopters earned roughly $400,000 more in FY25 than non-adopters, with the gap for large businesses running to $59.1 million. And that likely understates it, because the “adopters” bucket includes the confetti crowd alongside the genuine transformers.

The size divide is stark. MYOB’s April 2026 survey found firms with 100-plus employees reporting 37% improved profit margins from AI, versus just 11% for firms with 20 to 49 staff. MYOB executive Paul Voges described the situation as “the engine is firing on only half its cylinders.” MYOB’s August Business Monitor sharpened the point further, showing 92% of medium businesses would face immediate workload or cost hits if AI use stopped, against just 49% of small businesses. The firms going deep are becoming genuinely dependent. The ones dabbling could unplug tomorrow and barely notice.

Why the bill is about to force the issue

Datacom’s Peter Nelson flags a structural shift that will make shallow adoption impossible to ignore. The major platforms are moving from flat per-employee fees to consumption-based token pricing, which means the cost of vague, low-value use lands directly on the P&L. His framing is pointed: “You wouldn’t hire a 50-seater bus to transport two people from Hamilton to Wellington, so why use an expensive frontier AI model to improve an email?”

The fix is not more tools. The Ecosystm analysis from June 2026 identifies the problem as structural, arguing that real productivity “will come from redesigning work, data, roles, and governance rather than from cutting headcount or layering AI onto existing systems.” MYOB names five pillars: integrated processes, data systems, AI strategy, governance, and workforce capability. Awareness is not the barrier. MBIE research from 2025 found 94% of SMEs already know of at least one AI tool. Knowing where to start is.

What this means for the country

Prime Minister Christopher Luxon told the China Business Summit in June 2026 he was “constantly underwhelmed by how little we are adopting” AI, citing Estonia, Singapore and New South Wales as benchmarks New Zealand is missing. Compagnone notes that “in Australia, every single government department now has a chief AI officer.” New Zealand’s own strategy remains light-touch and principles-based.

The 91% figure will keep getting quoted as proof New Zealand is ahead of the curve. It proves the opposite. A country can have near-universal adoption and near-zero transformation at the same time, and right now that is exactly where we are. The businesses that redesign around AI will pull away. Everyone else is buying a slightly faster email at consumption-based prices.

Sources

• AI adoption hits 91% of NZ businesses, but use to transform core operations falls to 4% (2026-08-28)
• Most New Zealand businesses now use AI, so why has productivity gone backwards (2026-07-01)
• Most New Zealand businesses switched on AI, few use it well (2026-06-26)
• AI moves from hype to hard results for Kiwi businesses, 2degrees study reveals (2026-04-28)
• NZ’s mid-sized businesses yet to realise returns on AI investment: Survey (2026-04-24)
• New Zealand SMEs widen AI adoption gap, MYOB finds (2026-08-04)
• New Zealand’s AI Productivity Paradox: The Gap Between Deployment and Redesign (2026-06-23)
• Luxon ‘constantly underwhelmed’ by slow AI uptake in NZ businesses (2026-06-25)