# AI for New Zealand financial services. URL: https://levystreet.com/industries/financial-services/ Updated: 2026-10-19 How New Zealand banks, advisers, fund managers and lenders use AI for onboarding checks, policy search and reporting, and what the FMA and RBNZ expect. ## Short answer AI for financial services in New Zealand is most useful in document-heavy back-office work: checking client onboarding documents, answering staff questions from internal policies with citations, and drafting management reporting commentary. The FMA and Reserve Bank expect regulated entities to manage AI risk within existing obligations, and their October 2026 expectations say boards must know where AI is used and who is accountable. - RBNZ and the FMA set joint expectations on frontier-AI cyber risk on 15 October 2026. - The FMA's exploratory thematic review of AI in financial advice opened on 6 August 2026. - Internal policy retrieval with cited answers is a low-risk first project. - Lucent, a Levy Street venture, grounds investment research outputs in citations. ## Where AI fits in a regulated financial firm AI earns its place in financial services where staff spend hours reading documents and checking them against rules: onboarding files, policy manuals, product terms, reporting packs and research. It is a poor fit for the decisions regulators care most about, such as whether advice is suitable, whether a customer is treated fairly or whether a transaction is suspicious, unless a named person makes them. Interest is near universal among larger players. The FMA's [research published on 10 September 2024](https://www.fma.govt.nz/assets/Research/Understanding-Artificial-Intelligence-in-Financial-Services.pdf) surveyed 30 regulated entities and received 13 responses (six deposit takers, three insurers and four advice or asset management firms); all 13 had integrated generative AI or planned to soon. That is a small, self-selected sample, not a sector figure. The stakes are large: financial and insurance services contributed about $16.7 billion to Auckland's GDP and $3.8 billion to Wellington's in the March 2024 year ([Stats NZ regional GDP tables](https://www.stats.govt.nz/assets/Uploads/Regional-gross-domestic-product/Regional-gross-domestic-product-Year-ended-March-2025/Download-data/regional-gross-domestic-product-year-ended-march-2025.xlsx)). ## Financial services workflow map | Workflow | What AI does | What a person still decides | Data involved | Difficulty | |---|---|---|---|---| | [Client onboarding document checks](/industries/financial-services/client-onboarding-document-checks/) | Extracts identity, address and entity details, checks them against the onboarding checklist and flags expired or inconsistent documents | Sign-off of the onboarding file, the customer's risk category, enhanced checks | IDs, proof of address, trust deeds, company extracts | High: statutory AML/CFT duty | | [Internal policy retrieval](/industries/financial-services/internal-policy-retrieval/) | Answers staff questions from policies, procedures and product rules, quoting the clause it relied on | How the policy applies to a particular customer | Policy library, procedures, product terms | Low: read-only and cited | | [Management reporting commentary](/industries/financial-services/management-reporting-commentary/) | Drafts commentary on monthly results, arrears, flows or complaints, citing the figures | What the board is told, and risk ratings | Reporting packs, Power BI datasets, prior commentary | Medium: numbers must reconcile | | Adviser file notes | Drafts file notes and advice documents from meeting notes and the firm's template | The advice and its suitability for the client | Meeting notes, fact finds, risk profiles | High: advice duties | | Complaint triage | Classifies complaints, tracks internal dispute deadlines and drafts acknowledgements | Outcome, redress and any vulnerability response | Complaint emails, call notes, CRM | Medium: fair treatment at stake | | KYC refresh | Finds customers due for review and pre-fills what is already known | Whether information is current and adequate | CRM, previous due diligence | Medium: data quality varies | | Investment research synthesis | Summarises announcements, filings and broker notes with citations and dates | The investment view and any recommendation | NZX and ASX announcements, reports | Medium: dates and sources matter | | AI register upkeep | Keeps the inventory of AI systems, owners and controls current and drafts board updates | Risk appetite and acceptance | System list, incident logs | Low: structured records | Detailed guides cover [client onboarding document checks](/industries/financial-services/client-onboarding-document-checks/), [internal policy retrieval](/industries/financial-services/internal-policy-retrieval/) and [management reporting commentary](/industries/financial-services/management-reporting-commentary/). ## What the FMA, the Reserve Bank and other rules expect No regulator has written AI-specific rules for financial firms. Each has said existing obligations apply, and several have said more. - **FMA.** It calls itself [technology-neutral and pro-innovation](https://www.fma.govt.nz/news/all-releases/media-releases/understanding-artificial-intelligence-in-financial-services/) and expects most AI risks to be covered by its existing approach. On 6 August 2026 it opened an [exploratory thematic review of AI in financial advice](https://www.fma.govt.nz/library/reports-and-papers/thematic-review-ai-in-financial-advice/). At its April 2026 workshop with the University of Auckland's ALTeR centre there was [broad consensus](https://www.fma.govt.nz/assets/Reports/FMA-thematic-review-on-artificial-intelligence-in-financial-advice-engagement-overview.pdf) that primary responsibility for AI use rests with the advice licensee. - **Reserve Bank and FMA together.** In May 2025 the RBNZ said regulated entities must [manage AI risks within existing obligations](https://www.rbnz.govt.nz/hub/news/2025/05/rise-of-the-machines-how-could-artificial-intelligence-impact-financial-stability) and flagged reliance on a few AI providers as a concentration risk. On 15 October 2026 the two regulators set [joint supervisory expectations](https://www.rbnz.govt.nz/news-and-events/news/2026/10/rbnz-and-fma-outline-supervisory-expectations-on-cyber-and-ai-risks) on cyber risk from frontier AI: boards must know where AI is used, its risks and who is accountable. - **Fair conduct.** The Conduct of Financial Institutions regime [came into full effect on 31 March 2025](https://www.fma.govt.nz/news/all-releases/media-releases/cofi-regime-in-effect). Licensed banks, insurers and non-bank deposit takers must establish, implement and maintain a fair conduct programme, and AI in customer journeys sits inside it. - **AML/CFT.** The [AML/CFT Act](https://www.legislation.govt.nz/act/public/2009/35/en/latest/) puts the due diligence and record-keeping duties (ss 11 and 50) on the reporting entity itself. An onboarding system can collect and check documents faster; it cannot carry the obligation. This is general information. Talk to your compliance adviser or lawyer before changing a regulated process, and see our [AI governance advisory](/services/ai-governance/) for setting up the register and controls boards are now expected to oversee. ## Systems, data location and concentration risk Most firms run a core banking, registry or loan system, a CRM such as Salesforce, an advice platform, Microsoft 365 and Power BI. AI should read these through controlled interfaces, and our notes on [Salesforce](/integrations/salesforce/) and [Power BI](/integrations/power-bi/) cover the patterns. Data location needs a decision, not an assumption. As at October 2026 no major AI vendor offers guaranteed processing inside New Zealand; Amazon Bedrock's [AU profile](https://aws.amazon.com/blogs/machine-learning/run-generative-ai-inference-with-amazon-bedrock-in-asia-pacific-new-zealand/) keeps Claude processing within Australia and New Zealand. Given the Reserve Bank's concentration warning, design so you can switch models without rebuilding. Our guide to [data residency for AI systems](/engineering/data-residency-for-ai-systems/) sets out the options. ## A first project: internal policy retrieval with citations Policy retrieval is read-only, touches no customer decision and is easy to test, which makes it the right first project for most firms. - **Inputs**: The approved policy and procedure library, with version numbers and owners. - **Systems**: SharePoint or your document store, a retrieval index that respects access rights, and a model chosen by testing. - **Steps**: Staff ask a question; the system retrieves current clauses, answers with quotations and links, or says the policies do not cover it. - **Human review**: The compliance owner approves a 50-question test set before launch and reviews a weekly sample. - **Outputs**: A cited answer, or a referral to compliance. - **Measures**: Accuracy and citation correctness on the test set, referral rate, and time staff wait for an answer. - **Risks**: Superseded versions, answers that stray into advice, and questions outside the library. Illustrative example: a 60-person advice and fund management business fields about 40 policy questions a week, each taking a compliance officer 10 minutes, about 6.7 hours. If the assistant answers 60% with correct citations and refers the rest, compliance time falls by roughly 4 hours a week, around 200 hours a year. Launch only when the system scores at least 48 of 50 on the test set and refuses every out-of-scope question. [RAG development](/services/rag-development/) is how we build it. ## What Lucent shows about cited research [Lucent](/portfolio/lucent/) is a Levy Street venture: an agentic investment research platform for wealth, equity and private-markets teams, covering NZX and ASX, that runs several frontier models together and grounds outputs in citations. The same principle applies to any financial AI system: if a reviewer cannot click from a sentence to its source, the sentence should not be there. ## Where AI fails in financial services - **Stale documents.** Answers drawn from superseded policies or product terms. Index only current versions and show the version in each answer. - **Advice by drift.** A policy assistant starts telling staff what to recommend. Scope it to policy wording and refuse suitability questions. - **Numbers that do not reconcile.** Commentary quoting figures the model computed itself. Calculate every number in code from the dataset; let the model write only the words. - **Fake identity documents.** Generated IDs and altered statements are getting better. AI can flag anomalies; verification stays with approved methods and a person. - **Missed vulnerability.** Automated complaint replies overlook signs of hardship or distress. Route any such cue to a person. - **Single-provider dependence.** An outage stops the process. Keep a tested fallback model and a manual path. ## Questions people ask ### Does the FMA allow AI to be used in financial advice? There is no AI-specific rule for financial advice as at 19 October 2026. The FMA describes itself as technology-neutral and expects most AI risks to be covered by its existing approach. In August 2026 it opened an exploratory thematic review of AI in financial advice, and a stakeholder workshop it ran found broad consensus that the licensed advice provider bears primary responsibility for how AI is used. ### What do the RBNZ and FMA's October 2026 AI expectations require? They build on existing prudential, conduct, cyber and operational resilience obligations rather than adding a new rulebook. Boards and senior management need to know where AI is used in the business, understand its risks and be clear about who is accountable. Entities should also assume AI-enabled cyber incidents will happen and test their response plans against severe but plausible scenarios. ### Can AI carry out customer due diligence for us? AI can read identity and entity documents, compare them with your onboarding checklist, flag expiries, mismatches and missing beneficial owners, and prepare the file. The AML/CFT Act places the due diligence obligation on the reporting entity, so a trained person decides whether checks are complete, sets the risk rating, approves enhanced due diligence and handles any suspicious activity report. ### Where is our data processed if we use ChatGPT, Claude or Copilot? Usually outside New Zealand. As at October 2026 no major AI vendor offers guaranteed in-New Zealand model processing. Amazon Bedrock's AU profile keeps Claude processing within Australia and New Zealand, OpenAI's Australian region stores data there but processes it elsewhere, and Microsoft has not said where Copilot prompts from NZ tenants are processed. Record the location for each system you approve. ### How should we report AI use to the board? Keep an AI register: each system, what it does, the data it touches, where it is processed, the accountable owner, the human review step, the fallback if the provider fails, and incidents to date. That answers the RBNZ and FMA expectation that boards know where AI is used and who is accountable, and it makes concentration on a single provider visible. ## Sources - [Understanding Artificial Intelligence in Financial Services](https://www.fma.govt.nz/assets/Research/Understanding-Artificial-Intelligence-in-Financial-Services.pdf) - [Understanding artificial intelligence in financial services (media release)](https://www.fma.govt.nz/news/all-releases/media-releases/understanding-artificial-intelligence-in-financial-services/) - [Thematic review: AI in financial advice](https://www.fma.govt.nz/library/reports-and-papers/thematic-review-ai-in-financial-advice/) - [Thematic review on AI in financial advice: engagement overview](https://www.fma.govt.nz/assets/Reports/FMA-thematic-review-on-artificial-intelligence-in-financial-advice-engagement-overview.pdf) - [Rise of the machines: how could artificial intelligence impact financial stability](https://www.rbnz.govt.nz/hub/news/2025/05/rise-of-the-machines-how-could-artificial-intelligence-impact-financial-stability) - [RBNZ and FMA outline supervisory expectations on cyber and AI risks](https://www.rbnz.govt.nz/news-and-events/news/2026/10/rbnz-and-fma-outline-supervisory-expectations-on-cyber-and-ai-risks) - [Supervisory expectations on cyber risk associated with frontier AI (copy hosted by interest.co.nz)](https://www.interest.co.nz/sites/default/files/2026-10/Supervisory%20expectations%20on%20cyber%20risk%20associated%20with%20frontier%20AI.pdf) - [CoFI regime in effect](https://www.fma.govt.nz/news/all-releases/media-releases/cofi-regime-in-effect) - [Anti-Money Laundering and Countering Financing of Terrorism Act 2009](https://www.legislation.govt.nz/act/public/2009/35/en/latest/) - [Regional gross domestic product: Year ended March 2025 (data tables)](https://www.stats.govt.nz/assets/Uploads/Regional-gross-domestic-product/Regional-gross-domestic-product-Year-ended-March-2025/Download-data/regional-gross-domestic-product-year-ended-march-2025.xlsx) - [Run generative AI inference with Amazon Bedrock in Asia Pacific (New Zealand)](https://aws.amazon.com/blogs/machine-learning/run-generative-ai-inference-with-amazon-bedrock-in-asia-pacific-new-zealand/) Published by Levy Street (https://levystreet.com), Wellington, New Zealand.