This week in one paragraph
The AI conversation in New Zealand turned practical. A draft national AI policy was published that names small business as the priority user — the first substantive election-season proposal to do so. Every party's tech promises now have a live, continuously-updated tracker. Fresh survey data showed how shallow most AI adoption actually is, and the shortlist for the country's own AI awards went public. Promise, tracker, reality check, proof: that's the week, and none of it takes more than half an hour to act on.
The numbers on NZ AI use: 91% adoption, 4% transformation
What changed. Datacom's 2026 State of AI Index (Curia surveyed 207 senior leaders at organisations with 100+ staff, 28 July–3 August) found 91% of NZ organisations now use AI somewhere. But only 4% say it has transformed core operations — changed how the business actually runs — down from 8% a year ago, and just 15% are scaling AI across the whole organisation. Datacom's AI director, Lou Compagnone, calls the gap "AI confetti": staff using tools ad hoc while the business itself doesn't change. And 64% of surveyed businesses want NZ-specific AI legislation.
What it means. Adoption and transformation are different things, and the gap is wide: 81% of organisations are still in the exploratory or early-implementation stages, yet 79% increased AI investment over the past year. The money is moving faster than the strategy. For a small business, two things follow. First, watch costs: if a tool bills by usage rather than a flat fee, a "cheap" subscription can quietly climb as staff use it more. Second, don't mistake use for change — the firms getting value are the ones rebuilding a process around AI, not just switching it on.
Do this. Treat AI spend like any other cost: list your paid tools, note what each one returns, and check whether any bill climbs with usage.
Source: NZ Herald (premium link — paywalled) — https://www.nzherald.co.nz/business/ai-adoption-hits-91-of-nz-businesses-but-use-to-transform-core-operations-falls-to-4/premium/OL64JRBISVERTED3NQTN3PXHHA/
A draft AI policy aimed squarely at small business
What changed. Mark Laurence — who runs an AI training business, as he discloses in the piece — has published a draft national AI policy timed for the election. It proposes an Office of AI inside the Department of the Prime Minister and Cabinet, free AI literacy (knowing how to use AI day-to-day, not just knowing it exists) for every adult, split-cost vouchers so small businesses can afford proper help adopting AI, teacher upskilling, and Māori data-governance standards (rules for how Māori data is handled). The total ask is roughly $200–310M over three years — framed against the government's own AI strategy, which leans on a Microsoft estimate of $76B added to GDP by 2038 — and it calls on the next government to stand up a national AI conference within its first 100 days.
What it means. This is the first election-season proposal we've seen that treats small business as the priority user of AI, not a sideshow. The context is stark: 97% of NZ businesses have fewer than 20 staff, and Laurence says plainly that they are who this is for. Even if no party adopts it wholesale, the split-cost voucher idea is a direct, funded answer to the "we don't know where to start" blocker — and it gives you a concrete yardstick for testing every party's AI promises. One disclosure worth keeping in mind: Laurence flags a commercial interest in some of what he proposes, and asks that any funding go through open, contestable process.
Do this. Read the op-ed before you vote; it's a useful lens for judging what each party actually offers.
Source: interest.co.nz — https://www.interest.co.nz/technology/140260/mark-laurence-presents-his-proposed-draft-ai-policy-new-zealand-and-calls
The live guide to every party's tech promises
What changed. RNZ's Election 2026 technology policy guide is live and updated continuously as the campaign runs, and it covers every party with a realistic chance of entering Parliament — not just the big two. Labour has so far promised an AI office, an online safety regulator, copyright rules that pay creators, and rules on data centres.
What it means. These aren't abstract debates; they land on your business. An AI office shapes how government guidance and funding flow. Copyright rules touch the content you publish and the tools you use. Data-centre rules influence cloud and AI pricing in New Zealand. This is the single page that keeps a voter honest between now and election day.
Do this. Bookmark the guide and check it once a week; test each party's promise against the Story 2 yardstick.
Source: RNZ — https://www.rnz.co.nz/news/politics_election-2026/feature/technology-policy-guide
The shortlist for AI done well in NZ
What changed. AI Forum NZ announced the 2026 Aotearoa AI Awards finalists on 14 September across six award categories — including Te Ao Māori & Pasifika AI, Trailblazer in AI Innovation, Best Implementation of AI in Aotearoa, and AI Changemaker of the Year — plus the People's Choice Award, the supreme award chosen by public vote from all finalists. Voting is open until 11:59pm, 4 November. Winners are announced 18 November at AUT City Campus, Auckland.
What it means. The best-implementation finalists — a surveying firm, an online writing tool, structural engineering — are working examples of AI paying off in NZ trades and professions. That's the "here's what competitors may already be doing" list in concrete, local form: no Silicon Valley demos, just New Zealand businesses.
Do this. Spend 10 minutes browsing the finalists and voting in the People's Choice Award before 4 November — free benchmarking of what good looks like in Aotearoa.
Source: AI Forum NZ — https://aiforum.org.nz/2026/09/14/2026-aotearoa-ai-awards-finalists-announced/
One thing to do this week: the 15-minute AI spend audit
Open your banking app and app-store receipts. List every AI tool you or your staff pay for, note what each one actually gets used for, then check whether any bill is metered by usage and creeping up. Cut what isn't earning its keep. Then pick one business process and rebuild it with AI this quarter — that single decision is the difference between the 91% who use AI and the 4% whose operations have actually changed because of it.