A growing share of New Zealanders are now asking AI tools to do the job a broker traditionally does: compare policies, decode pricing, and work out what actually offers value.
The reasons why sit uncomfortably close to the industry's own recent history of enforcement action. New research from Accenture Song and fiftyfive5, surveying 2,364 New Zealanders across 42 retail brands between April 23 and May 31, 2026, shows this is not a fringe behaviour.
It also shows, in the same dataset, why customers may be looking for an alternative to asking their insurer directly.
Among general insurance customers surveyed, 11% said they had already used a third-party AI tool to help with a category decision, and 21% of that group said it left them more confident in the choice they made.
Across every category in the study, the most common reason people turned to AI was simplifying product or service comparisons, cited by 45% of respondents, followed by personalised recommendations at 37% and price optimisation at 34%.
That is not a small slice of the market, and it is a slice that historically has had one obvious human alternative: a broker. What is notable is that the same customers describing their reasons for reaching for AI are, almost word for word, describing the value proposition brokers are meant to provide.
One respondent said the industry needed to "simplify the insurance and make it easy to understand," describing their own policy as "a minefield of information that almost needs decoding - and then the price keeps going up until it becomes unaffordable."
Another wanted someone to walk them through their cover annually rather than hearing from their insurer only at renewal.
Some of that frustration has a documented history behind it. In October 2024, the Auckland High Court ordered AA Insurance to pay $6.175 million after it admitted failing to apply multi-policy discounts, NZAA membership discounts and guaranteed no-claims bonuses to well over 100,000 customers, resulting in roughly $11.12 million in overcharges.
Justice Laura O'Gorman's judgment noted: "Customers cannot be expected to double check the precise details of transactions. They are entitled to trust the accuracy of their insurer's systems and processes." A year later, in October 2025, IAG New Zealand received an even larger penalty - $19.5 million, the biggest fair dealing case the FMA has brought - again over premium pricing and discounts.
Those cases sit inside a wider climb in dispute volumes. Insurer disputes lodged with Financial Services Complaints Limited rose roughly 39% over two consecutive years, from 46 cases in 2023/24 to 64 in 2025/26, out of 1,596 complaints in the latest twelve-month period. The Insurance and Financial Services Ombudsman Scheme accepted a record 600 disputes for investigation in 2024-25, more than double its 2022 total, with delays and poor communication - not disagreements over coverage - the consistent driver. The Financial Markets Authority's latest Financial Conduct Report puts nearly $29.8 million in penalties and enforceable undertakings against insurers over the past year, with general insurance making up 67% of complaints and travel insurance alone accounting for 18%.
If customers are drifting toward AI because the direct insurer relationship has let them down, that ought to open ground for brokers rather than close it. Except the same distrust dynamic is now appearing inside the AI tools themselves.
The Insurance and Financial Services Ombudsman Scheme has warned that AI tools some consumers use to draft complaints or estimate dispute outcomes have produced incorrect legal interpretations, invented case references and misleading claims about the odds of success. A client acting on a chatbot's confident but wrong reading of their own policy is not obviously a better outcome than a client who never got advice at all - and it raises the stakes for the person actually accountable for what a client understood before a claim goes wrong.
There is also a policy dimension building in the background. Deloitte's latest outlook on the New Zealand market notes that the government has launched a review, led by the Council of Financial Regulators, into the impact of rising premiums on residential insurance uptake and underinsurance. A review built around pricing transparency and affordability sits close to the exact terrain - explaining what a policy costs and why - that both AI tools and brokers are now competing to occupy.
General insurance's overall experience gap - the difference between what the sector promises and what customers say they receive - sat at 69% in 2026, down two points from 71% the year before. Progress is real, but it has not closed fast enough to stop customers from finding their own workarounds.
For brokers, the data describes a specific competitive moment. A broker who reviews a client's policy wording annually, explains what it covers and what it does not before a claim arrives, and presents that explanation in plain language is doing something a comparison AI tool structurally cannot do - because AI tools work from what is publicly stated, not from a working knowledge of the wording, the client's actual situation, and the difference between them. Whether that gap gets reclaimed by brokers, or continues to be filled by AI tools that are sometimes wrong and always unaccountable, will depend on whether brokers make that service visible before a client reaches for the chatbot rather than after.
