Half of New Zealand businesses are now consistently using artificial intelligence, but only 15 percent have a formal AI strategy in place, according to new research commissioned by Amazon Web Services.
That difference may be becoming more important than the adoption figure itself. As businesses move beyond testing individual AI tools, decisions around where AI is used, how projects are funded, who is responsible, and how returns are measured are starting to separate experimentation from wider deployment.
AWS’s Unlocking New Zealand’s AI Potential 2026 surveyed 1,000 businesses and 1,000 individuals. It found 50 percent of businesses were consistently using AI, up from 42 percent a year earlier.
Among AI-adopting businesses, 84 percent reported an increase in revenue, while 32 percent had launched a new AI-driven product or service during the past year. A further 27 percent said AI had helped them expand internationally.
Those figures point to businesses finding commercial applications for the technology, although the survey does not establish how much of the reported revenue growth can be attributed directly to AI.
The sharper distinction was between businesses with and without a formal strategy.
Among businesses with an AI strategy, 41 percent had reached what AWS classified as advanced AI use, compared with 19 percent of other businesses. Some 63 percent said they were confident they could measure return on investment, against 34 percent without a strategy.
The gap was not confined to large organisations. AWS found startups were ahead of both large enterprises and SMEs in advanced AI use, suggesting access to capital or scale alone was not determining how quickly businesses progressed.
For manufacturers and other operational businesses, the more relevant question is increasingly where AI can remove cost, reduce lead times, support product development, or take repetitive work out of existing processes.
New Zealand companies are already testing those applications.
ZURU is using AI to analyse around 20,000 social videos a day, with AWS reporting that the company has reduced its product development cycle from 18 months to five months. Agritech company Halter developed an AI agent to help triage system alerts, which it said had saved more than 215 hours of engineering time.
Parrot Analytics is using Amazon Bedrock to verify content data, reporting a tenfold increase in processing speed and a 60 percent reduction in cost. Tower Insurance, working with Deloitte, deployed an AI-powered contact centre in 10 weeks and reported saving more than one million minutes of customer call time.
The research also highlighted a second constraint: skills.
More than half of New Zealand individuals surveyed, 54 percent, said they had used an AI tool to teach themselves a new skill during the past year. Only 34 percent said they had received AI or digital training through their employer.
AWS identified a lack of skilled personnel as the leading barrier to adoption. The research put the average time required to fill a digitally skilled position at 3.4 months, while 52 percent of new roles were expected to require AI literacy within three years.
That creates a different issue for businesses already investing in technology. Buying access to AI is relatively straightforward. Building enough capability across a workforce to identify useful applications, assess the output, manage risk, and measure whether a project is actually improving the business is harder.
The research coincides with the first anniversary of the AWS Asia Pacific New Zealand Region. AWS said thousands of customers were now running workloads through the local Region, with organisations including Foodstuffs North Island, Halter, Xero, ZURU, Air New Zealand, and Ravensdown among those working with the company in New Zealand.
AWS is also extending its training programmes, including a new AI Business Strategist certification aimed at people responsible for AI investment, business cases, and adoption rather than technical development.
For businesses, the next stage of AI adoption may therefore be less about adding another tool and more about deciding which applications deserve to move beyond trial.
With only 15 percent of businesses reporting a formal strategy, there is still a sizeable gap between widespread interest in AI and having the management structure to decide what is worth scaling.
Read more here.
