In 2026, 91% of New Zealand organisations use AI in some form — but only 4% have used it to transform how their core business runs, and just 36% of SMEs use it proactively. The gap between “we use ChatGPT” and “AI does real work for us” is the whole story. The businesses closing it are starting with one costly, repetitive process, keeping data local and compliant, and measuring the result.
Two years ago, most conversations about AI in New Zealand were curiosity. Now the surveys are in, and they say something more interesting than “everyone’s using it”. Almost everyone is using it — and almost nobody is using it well yet. This post pulls together the 2026 New Zealand data from Datacom, MYOB, Deloitte/2degrees and the AI Forum, and explains what it means if you run a business here.
Adoption is near-universal — depth is not
Datacom’s 2026 State of AI Index puts AI use at 91% of New Zealand organisations, up from 87% in 2025 and 66% in 2024. But 81% are still in the exploratory or implementation stages. Only 15% are scaling AI organisation-wide, and the share using AI to transform core operations actually fell — from 8% to 4%.
The type of AI in use explains the gap. In the same survey, 68% use general-purpose assistants (ChatGPT, Gemini, Claude) and 66% use AI features inside productivity tools. Far fewer have gone further: 31% use AI embedded in business platforms, 26% have custom-built or in-house AI, and just 13% run agentic systems — AI that takes actions across systems rather than answering questions.
- 91% use AI in some form (Datacom, 2026)
- 15% are scaling it organisation-wide; 4% have transformed core operations
- 26% have custom-built AI; 13% use AI agents
- 62% have some form of AI strategy, but only 16% a clearly defined standalone one
Small businesses are further behind — and it depends on size
The MYOB Business Monitor (1,026 NZ business owners and managers, August 2026) found 36% of SMEs proactively using AI, up only four points on the year before. Size is the dividing line: 30% of businesses with 1–5 staff use AI, against 64% of those with 20 or more. And where it is used, it has become load-bearing — 92% of those medium businesses say stopping AI use would hurt workload, productivity or costs.
What SMEs use it for is telling: social media and marketing content (38%), marketing copy (28%), documents and reports (26%), and market analysis (25%). That is content generation, not operations. The invoices, enquiries, bookings and job admin that consume most of a small business’s week are largely untouched.
The return is real for those who go further
Deloitte Access Economics, in research commissioned by 2degrees (April 2026), found that NZ SMEs using AI earned around $400,000 more in FY25 than comparable non-adopters, and that AI now accounts for 29% of business technology spend, heading for 34% within two years. Datacom’s index adds that 79% of organisations increased AI investment in the past year, 33% say returns already exceed costs, and the most-cited benefit is productivity and time saved (59%), followed by quality and accuracy (41%).
That matches what we see on the ground: the projects that pay back are not the ones with the most impressive demo. They are the ones pointed at a process that costs real hours every week — customer enquiries, invoice and document processing, quoting and follow-up, and turning scattered data into dashboards.
What is holding businesses back
The barriers have shifted. In Datacom’s 2026 survey, lack of internal capability is still the top barrier at 20% — but it fell 12 points in a year. Rising instead: employee concern about job displacement (18%, up 7 points) and implementation cost (17%, up 8 points). The AI Forum NZ estimates only 2.7% of the workforce are AI practitioners and 79% of employers are unsure how to train staff — which is why 66% of organisations ran AI skills training last year, up 10 points.
In practice, most stalled projects we are asked to rescue share one of three patterns: activity without an owner or a metric; a tool bought before the process was defined; or a compliance question — usually about where data goes — that nobody answered before launch.
The New Zealand angle: keep it local and compliant
Where your information is stored and how it is handled matters, both for customer trust and for the Privacy Act 2020. The businesses adopting well insist on New Zealand or Australian hosting where the platform allows, send personal information to a model only when the task needs it, use enterprise terms that exclude training, and build audit logs in from the start rather than bolting them on later. Datacom’s index found 64% of organisations want specific NZ AI legislation and 74% support a national AI framework — expect the compliance bar to rise, not fall.
The organisations getting value from AI aren’t chasing the technology. They’re solving one real problem at a time, keeping their data safe, and expanding from there.
What this means if you run a business in New Zealand
- Using ChatGPT is table stakes, not an advantage. Nine in ten organisations already do. The 4% advantage is in operations.
- Start with a process, not a tool. Pick the one repetitive job that costs the most hours — enquiries, invoices, bookings, follow-ups — and scope an AI agent or automation around it.
- Agree the metric before you build. Hours saved, response time, resolution rate. The 33% seeing returns above cost are the ones who measured.
- Answer the data question first. Where it lives, who can see it, what the model receives. It is cheaper to design in than to retrofit.
- Bring your team along. Displacement worry is the fastest-growing barrier; agents that hand off to people with full context get adopted, agents that replace people quietly get sabotaged.
You don’t need a grand AI strategy to start — you need one clear, costly problem and a small, well-scoped first project. That’s how our AI business solutions work: fixed-price, NZ-based, and designed to prove value before you scale. The gap between the businesses using AI well and those still watching is widening, and it’s a smaller step across than most people think.
Sources
- Datacom, 2026 State of AI Index (New Zealand organisations; 2026)
- MYOB Business Monitor, August 2026 — reported by IT Brief NZ (n = 1,026 NZ business owners, managers and directors)
- Deloitte Access Economics for 2degrees, April 2026 — reported by IT Brief NZ
- AI Forum New Zealand workforce and adoption figures, 2026 — via Cairn
Last updated 15 September 2026. Written by Ashok Poshamalla, Founder, Avtrix Software Solutions, Taupō.