The Financial Markets Authority launched a thematic review on 6 August into how firms use artificial intelligence in financial advice, sending four surveys to the industry.
They go to advice firms, technology suppliers, lawyers and industry bodies, and are open until 4 September 2026.
The firm survey asks about "governance, oversight, suitability, record-keeping and consumer outcomes". In plain words, who is in charge of the AI, who checks its work, and whether customers end up better or worse off.
One answer firms can tick in the advisers survey is that "AI outputs are used directly with customer without adviser review".
Nobody is in trouble. The FMA calls the work "an exploratory exercise rather than a formal review or enforcement initiative", in its engagement overview.
The same document says the existing framework, including the Code of Professional Conduct for Financial Advice Services 2025, "is technology-neutral".
It also records a workshop with University of Auckland researchers on 15 April 2026. More than 40 industry people raised "concerns about AI outputs blurring the distinction between the provision of factual information and giving regulated financial advice".
Question 34 of the firm survey asks which parts of the law leave firms most unsure about AI. One of sixteen options is the "distinction between regulated financial advice and other forms of guidance or information".
Question 5 splits "research and product comparison" from "suitability assessment/ recommendations", which is the same divide one step earlier.
The line comes from the Financial Markets Conduct Act 2013. Section 431C says recommending, or giving an opinion on, buying or selling a particular financial product is financial advice. Insurance contracts count, so life, health, trauma and income cover are all in scope.
Regulated advice carries duties. The adviser must be competent, put the client's interests ahead of their own, and work with care, diligence and skill.
Schedule 5 of the Act says "providing factual information", such as the cost or terms of a product, is not advice. Nor is a recommendation about a kind of product in general.
Neither rule mentions technology. By our reading, which isn't legal advice or the FMA's position, an AI tool that ranks named policies and badges one as right for you is giving advice. One that reads out a published exclusion isn't.
The FMA has published no guidance on where AI outputs fall. The overview says it "may" publish a high-level summary, talk to the industry further, or use what it learns to shape future guidance.
No date is set. Reports drawn from the questionnaire "will generally be aggregated and anonymised".



