Datacom's NZ AI Index: 91% use AI. 4% transformed core operations.

Adoption went up. The top of the maturity curve halved. You've probably briefed a board on usage while the operating model stayed exactly where it was last year. If that does not haunt your L&D dashboard, you are measuring the wrong thing.

Peter Nelson is Managing Director of Datacom New Zealand. Lou Compagnone is Director of AI. Their Index is the artefact. Not a vendor fairy tale about a single customer.

Ninety-one per cent of New Zealand organisations now use some form of AI. Four per cent say it has transformed core operations. That 4% is down from 8% a year earlier. Read it again. Adoption went from 66% in 2024 to 87% in 2025 to 91% now. The top of the maturity curve halved. Eighty-one per cent are still exploring or implementing. Fifteen per cent are scaling across the organisation, up from 12%. That is a country getting better at buying tools and worse at claiming the operating model moved.

Nelson said it is easier to introduce AI for basic everyday use than to scale it. Widespread use is not organisation-wide change. The shops making progress treat it as a business priority with clear ownership, clear objectives and a focus on where it creates value. Compagnone said too many still treat AI as a technology initiative bolted onto a CIO, CTO or chief data officer. It now hits strategy, workforce, governance, risk and change. The real gap is not users versus non-users. It is adopters versus people willing to redesign the work.

Leadership, skills, spend

Twenty-two per cent have a dedicated AI leadership role (13% a CAIO, 9% another full-time seat). Twenty-six per cent added it to an existing executive job. Nineteen per cent share it across several people. If nobody owns it, the 4% will not become 8% again.

The main barrier to scaling is lack of internal capability or skills (20%). Then employee concern, resistance or fear of displacement (18%). Then implementation cost (17%). They are trying: 59% employ or plan to employ dedicated AI talent, up 25 points. Sixty-six per cent provided AI skills training in the past 12 months, up 10 points. Training is up. Core-operations transformation is down. That should haunt every L&D dashboard in Wellington.

Seventy-nine per cent increased AI investment last year. Seventy-three per cent expect to spend more next year. A third say returns exceeded costs. Another 30% say they roughly broke even. Value they cite: time and productivity (59%), quality and accuracy (41%), operational efficiency (34%). Tools: general-purpose assistants 68%, productivity add-ons 66%, embedded in platforms 31%, custom 26%, agentic 13%. Strategy: 62% have some form. Only 16% have a clear standalone strategy. One in five have not addressed cost management. Only 20% have a dedicated AI budget with tracking.

On policy, 55% are somewhat or very confident New Zealand can manage the risks and opportunities (up 18 points). Sixty-four per cent want specific AI legislation. Seventy-four per cent want a national framework and a coordinating body. Nelson wants clearer rules as AI gets closer to critical services.

If you run a New Zealand enterprise, this Index is your artefact. Do not wait for an Australian bank's ASX pack. Skills work that redesigns a workflow is the only training number that can move that 4%. Completions will not.

Look at the tool mix before you buy another platform. Sixty-eight per cent are on general-purpose assistants. Sixty-six per cent on productivity add-ons. Thirteen per cent on agentic systems. The country is still in ChatGPT-and-copilot land. The 4% who transformed core operations are not hiding in the 13%. They are hiding in the operating-model sentence Compagnone keeps repeating.

Cost discipline is the sleeper. One in five have not addressed AI cost at all. Only 20% have a dedicated budget with tracking. You can raise spend (79% already did, 73% will again) and still not know whether the third who say returns beat costs are telling the truth. Nelson wanting clearer national rules is downstream of that. Boards hate ambiguity. They will still sign the invoice.

The leadership chart is a hiring brief. Thirteen per cent have a CAIO. Another 9% have a full-time AI lead. The rest added it to the CIO or formed a committee. If you are in the 26% or the 19%, you are who Compagnone is talking about when he says technology initiative. Give it an owner or stay in the 81% still exploring.

What this means for L&D

The only slide you need is 91 and 4. Name Nelson and Compagnone. If skills are the top barrier and training is already up 10 points, change the training, not the volume. Completions moved. The operating model did not.

Sources: IT Brief NZ, 2 September 2026.