New Zealand businesses may need to spend more than NZD $300 million over five years to stay visible to online consumers as AI tools change shopping behaviour. The estimate comes as survey data suggests AI is starting to displace conventional search engines for some shoppers.
Research cited by Sprocket Digital found 44% of New Zealand consumers who have used AI while shopping have reduced their use of traditional search engines. The same survey found 64% use AI to compare products and prices, while 62% use it to research planned purchases.
Jono Swanston, General Manager at Sprocket Digital, said the shift could leave domestic companies at a disadvantage if larger overseas rivals become easier for AI systems to identify and recommend. He said the response would require changes to websites, content and broader digital strategy, rather than simple adjustments to search rankings.
"Historically, someone might go to Google and search for something very simple such as 'emergency plumber Auckland'. In contrast, AI search is prompt-oriented, so that same person might describe exactly what type of plumber they need, where they are, what type of property they have and what problem they are trying to solve.
"That means businesses need enough relevant information online for an AI platform to understand not simply what they do, but where their expertise lies and when they should be recommended over a competitor.
"The concern for New Zealand is that consumers are moving in this direction faster than many businesses are preparing for it," said Jono Swanston, General Manager at Sprocket Digital.
The argument reflects a broader concern in digital marketing that generative AI systems are changing the path consumers take before making a purchase. Instead of scanning a page of search results, users can ask a detailed question and receive a summarised recommendation or comparison drawn from multiple sources.
That matters for smaller markets such as New Zealand because location offers less protection when consumers ask AI tools to compare suppliers globally. Exporters and online retailers may find themselves competing more directly with international marketplaces if those services are more widely referenced and better understood by AI models.
"Geography provides less protection when AI is comparing businesses globally. New Zealand companies are already competing online with very large international marketplaces and businesses with much greater resources. If those overseas companies are better understood and more visible to AI platforms, local businesses could simply be left out of the answer.
"That has the potential to translate into a significant loss of revenue for New Zealand businesses and more consumer spending moving offshore," Swanston said.
Website changes
Swanston said many existing websites would not need a full rebuild, but they may need technical updates and more detailed content. In his view, businesses need to publish clearer information about what they do, who they serve, where they operate and the specific problems they solve.
"Most existing sites will not need to be completely rebuilt, but may require technical changes to make information easier for AI systems to access, along with substantially richer content designed around the questions customers are now asking.
"You can retrofit an existing website. There are technical components in the backend, but a major part of it is the quality and structure of the content.
"The old approach of putting a particular keyword in the right places and trying to rank number one is no longer enough.
"Businesses need to answer a much broader range of questions and make it very clear what they do, who they do it for, where they operate and what makes them relevant to a particular customer need," Swanston said.
Media role
The research also points to a potential shift in the value of earned media. Analysis of more than 25 million links cited by ChatGPT, Claude and Gemini across 17 industries found 84% of AI citations came from earned media, with journalism accounting for 27% of cited sources. Paid and advertorial material made up 0.3%.
That finding suggests AI systems may rely heavily on independently produced reporting when forming answers about companies, industries and products. It also raises the possibility that news coverage could influence which businesses are surfaced in AI-generated recommendations.
"Businesses have always been able to make claims about themselves on their own websites, but an independent source saying something about that business provides another level of validation.
"News coverage can give AI systems context about what an organisation is known for, which industry it operates in, who its experts are and whether its claims are supported by independent sources.
"That means the value of a credible media article may increasingly extend well beyond the people who read it when it is first published," Swanston said.
The same analysis found journalism citations were spread across more than 20,000 outlets rather than concentrated among a small number of global publishers. Industry trend questions also generated more than twice as many journalism citations as how-to questions, indicating a particular role for specialist and topical reporting.
Swanston said that trend could increase the relevance of national, regional and trade publications as AI tools become a common way for consumers to research firms and sectors. At the same time, he said publishers face a familiar problem as AI systems answer questions directly without always sending users to the original source.
"We're moving into an environment where information can be taken from a website and incorporated into an answer without necessarily generating the website visit that would historically have followed.
"So news media could become more influential as an information source while at the same time having to work out how to capture the commercial value of that influence," Swanston said.
Skills gap
Another issue is whether New Zealand has enough people with the right mix of search, content and technical knowledge to help firms adapt. Swanston said the market may struggle if demand for this kind of work rises quickly, because AI search practices are changing too fast for training pipelines to keep pace.
He said companies can start with a basic test by asking AI platforms the kinds of questions their customers would ask, then checking which businesses are recommended and how accurately they are described. That can indicate whether a company has an AI visibility problem before it commits significant spending.
The issue may become more pressing if consumers move beyond using AI for research and allow software agents to complete purchases. Survey data from PwC found 31% of consumers in New Zealand were already comfortable letting an AI agent buy on their behalf.
"Today the question is whether ChatGPT recommends your product or your business. The next phase is potentially an AI agent comparing the options, making the selection and completing the transaction for the customer.
"At that point, being invisible to AI could mean being excluded from the purchasing process altogether.
"This is why New Zealand businesses need to treat AI search as more than another SEO update. It is a structural change in how consumers find, compare and ultimately choose who they buy from," Swanston said.