AI in Finance in Aotearoa New Zealand: A Living Whitepaper

Snapshot date: July 13, 2026
Updated from the June 10, 2026 edition

Introduction

AI adoption in New Zealand finance is still best described as practical, supervised, and infrastructure-dependent. The strongest live use cases remain fraud and scam prevention, customer-service augmentation, adviser productivity, workflow automation, and the open-banking rails that make more advanced data-driven services possible. What has become clearer since the last edition is that firms are moving from isolated pilots toward named executive ownership, measurable frontline outcomes, and broader ecosystem integration. (westpac.co.nz)

Executive Summary

• The dominant model is still “AI with humans in charge.” The FMA’s latest advice-sector work says AI is mostly being used to enable rather than replace advisers, while Westpac’s contact-centre rollout and Tower’s insurer contact-centre deployment both keep staff in the loop. (fma.govt.nz)
• The most important post–June 10 shift is operational maturity. Westpac now has a dedicated Chief Data, Digital and AI Officer, signalling that AI is becoming a formal executive portfolio rather than a side programme. (westpac.co.nz)
• Open banking is becoming the key enabling layer for the next phase of finance AI. Kiwibank’s rollout is live, API Centre reporting shows it marked “READY” on 28 May 2026, all four largest banks have partnered with at least one third party on the common payment standard, and live use cases now include instant account top-ups, merchant payments, and peer-to-peer-style payment requests. (kiwibank.co.nz)
• Measured customer-service gains are now public. Tower says its AI-enabled contact centre saved customers more than 796,000 minutes over seven months, cut interaction time by about 15%, and reduced average handling time by 2 minutes 38 seconds. (tower.co.nz)
• Risk language remains firm and is broadening. RBNZ continues to warn that AI can amplify concentration, model, cyber, and even credit risk, and its July 8, 2026 Monetary Policy Review also noted the possibility of a correction in AI-related asset prices affecting financial conditions. (rbnz.govt.nz)
• The market is still advancing faster on defensive and assistive use cases than on autonomous decisioning. Fraud disruption, scam detection, service-assist tools, and adviser workflow compression remain the strongest proof points. (anz.com.au)

What’s New Since June 10, 2026

1) Westpac has elevated AI to a dedicated executive function

Westpac’s current executive team listing shows Russell Jones was appointed Chief Data, Digital and AI Officer in June 2026, with responsibility for setting and leading Westpac New Zealand’s data, digital and AI strategy. That is a meaningful governance signal: AI is no longer just a technology capability inside the CIO function, but part of named executive accountability. (westpac.co.nz)

2) Insurance has added one of the clearest public outcome-based AI case studies

Tower’s May 19, 2026 update is one of the strongest public New Zealand finance examples of AI at operating scale. Tower said its AI-enabled contact centre saved customers more than 796,000 minutes in seven months, reduced time spent interacting with the insurer by around 15%, and lowered average handling time by 2 minutes 38 seconds per interaction. It also said real-time transcription now runs across sales, service, and claims calls. (tower.co.nz)

This matters because it moves the sector’s evidence base beyond bank pilots and adviser tools. It shows AI producing measurable service and productivity effects inside a regulated insurance workflow, while still being used as staff support rather than autonomous adjudication. That is an inference from Tower’s deployment description. (tower.co.nz)

3) Open banking is now looking less like policy plumbing and more like usable market infrastructure

API Centre implementation reporting now shows Kiwibank marked “READY” on 28 May 2026, and Kiwibank’s own release says it rolled out open banking across all digital channels for personal and business customers on that date. MBIE’s regime remains in force, with Kiwibank required to have payment services ready from June 2026 and account-information services ready from December 2026. (apicentre.paymentsnz.co.nz)

The ecosystem is also getting richer. API Centre says all four largest banks have now partnered with at least one third party on the standardised payment-initiation API, while current live examples include BlinkPay’s integration with Sharesies for instant top-ups, Qippay merchant payments, Volley payment requests, and Worldline’s Online EFTPOS, which it says is live with more than 500 merchants and has been used by around 700,000 New Zealanders. (apicentre.paymentsnz.co.nz)

For AI in finance, this is strategically important because consented data-sharing and payment-initiation rails are prerequisites for more intelligent cashflow tools, embedded finance, personalised recommendations, and automated financial workflows. That is an inference from the open-banking rollout evidence. (apicentre.paymentsnz.co.nz)

4) Real-time payments capability is now being publicly demonstrated

On May 27, 2026, BNZ said BlinkPay had proven real-time payments could work in New Zealand using open banking, with BNZ as the first bank to participate in testing. Test payments appeared in a recipient BNZ account in 2.6 seconds from customer approval to funds arrival. (bnz.co.nz)

That is not AI adoption by itself, but it strengthens the transaction rails that AI-enabled finance products can sit on top of. Faster payment confirmation can materially improve the usefulness of AI-assisted treasury, merchant, lending, and customer-experience workflows. This is an inference from the BNZ/BlinkPay result. (bnz.co.nz)

Current State of AI Adoption in New Zealand Finance

Sector-wide position

The FMA’s September 2024 baseline remains the clearest broad market starting point: all 13 participating firms across banking, insurance, asset management, and financial advice either already used generative AI or expected to adopt it soon. Its March 2026 access-to-advice work adds that the predominant use of AI in advice is to support advisers rather than replace them. (fma.govt.nz)

The best current synthesis is that New Zealand finance has moved beyond experimentation, but not into widespread autonomous consequential decision-making. AI is increasingly embedded in workflows, fraud systems, customer operations, and advice support, with human accountability still central. That is an inference from the combined regulator and firm evidence. (fma.govt.nz)

Where adoption is strongest

• Fraud, scams, and cyber defence: ANZ said its bank-telco-tech coalition disrupted more than 5,000 phishing domains in two months and helped reduce ANZ customer card-phishing cases by 39% over the same period, while warning that scams are becoming more complex and AI-generated attacks are rising. (anz.com.au)
• Customer-service augmentation: Westpac’s Microsoft contact-centre deployment and Tower’s AI-enabled contact centre are the strongest current public examples of AI assisting frontline staff in regulated interactions. (westpac.co.nz)
• Advice productivity and hybrid advice: The FMA says AI-supported advice processes could improve scalability and consistency with proper oversight, and Deloitte’s NZHL case study says annual client-review preparation time was reduced by 80%. Because NZHL is a vendor-published case study, it is best treated as directional rather than independent evidence. (fma.govt.nz)
• Open-banking-enabled payments and data sharing: Kiwibank’s rollout, API Centre implementation progress, and BNZ’s real-time payment testing all show that the data and payments layer underpinning future AI services is maturing quickly. (kiwibank.co.nz)

Research and Policy Overview

What regulators are saying now

The FMA’s March 2026 advice review is still the most useful public document on how AI is being applied inside a consequential financial workflow. It identifies AI agents, compliance-checking tools, client-analysis tools, note summarisation, and proactive financial-wellbeing tools as examples already visible in the market, while also stating that the FMA plans a thematic review of the use of AI in financial advice. (fma.govt.nz)

Consumer trust is present but conditional. In the FMA’s consumer research, net trust in AI-provided advice ranged from 28% to 41% depending on product type; 82% preferred face-to-face interaction with an adviser, and only 3% preferred AI as the interaction method. (fma.govt.nz)

RBNZ’s May 2026 Financial Stability Report remains the key prudential reference point. It warns that relying on only a small number of third-party AI providers could create dependencies, increase the risk of biased or fraudulent outputs, and amplify cyber threats; it also notes a possible credit channel if AI materially affects employment in some sectors. (rbnz.govt.nz)

What changed in July

RBNZ’s July 8, 2026 Monetary Policy Review added another useful macro signal: global growth has been resilient partly because of strong AI-related investment, but the Committee also noted the risk of a correction in AI-related asset prices, with implications for financial conditions and stability. That does not change the local adoption picture directly, but it reinforces that AI now matters to New Zealand finance both as an operational technology and as a macro-financial transmission channel. (rbnz.govt.nz)

Case Studies

Case Study 1: Westpac NZ’s AI-supported contact centre

What happened: Westpac began rolling out Microsoft Dynamics 365 Contact Centre as a Service with built-in AI on April 14, 2026, aiming for deployment across all contact centres by August 2026. The system surfaces relevant customer and product information in real time to support staff. (westpac.co.nz)

Why it matters:

• It is one of the clearest New Zealand banking examples of AI in live regulated customer interactions. (westpac.co.nz)
• Westpac has now paired that deployment with dedicated executive AI ownership through its June 2026 Chief Data, Digital and AI Officer appointment. (westpac.co.nz)

Case Study 2: Tower’s AI-enabled contact centre

What happened: Tower says its AI-enabled contact centre reduced customer interaction time by around 15%, lowered average handling time by 2 minutes 38 seconds, and saved more than 796,000 minutes over seven months. (tower.co.nz)

Why it matters:

• It is one of the best public outcome-based AI deployments in New Zealand financial services. (tower.co.nz)
• It shows AI being embedded directly into everyday claims and service workflows, not just back-office experimentation. (tower.co.nz)

Case Study 3: Kiwibank’s all-channel open-banking rollout

What happened: Kiwibank says it rolled out open banking through all digital channels for personal and business customers on May 28, 2026, and API Centre reporting shows it as “READY” from that date. (kiwibank.co.nz)

Why it matters:

• It turns open banking from timetable into customer-usable infrastructure. (kiwibank.co.nz)
• For AI, it expands the trusted consent and action rails needed for higher-value automation and personalisation. This is an inference. (westpac.co.nz)

Case Study 4: BNZ and BlinkPay’s real-time payments proof

What happened: BNZ said BlinkPay demonstrated open-banking-based payments reaching a recipient BNZ account in 2.6 seconds. (bnz.co.nz)

Why it matters:

• It shows the payments layer is becoming fast enough for genuinely real-time financial workflows. (bnz.co.nz)
• That improves the practicality of AI-assisted merchant, treasury, and customer-service experiences. This is an inference. (bnz.co.nz)

Key Trends

Trend 1: Defensive AI still leads

Scam detection, phishing disruption, fraud monitoring, and identity-security tooling remain the strongest proven AI-adjacent use cases in New Zealand finance. (anz.com.au)

Trend 2: Customer-facing AI is scaling where accountability is clear

Banks and insurers are expanding AI in contact centres and service operations, but the visible pattern is augmentation rather than replacement. (westpac.co.nz)

Trend 3: Open banking is becoming the sector’s AI substrate

The more mature the consented data-sharing and payment-initiation layer becomes, the easier it is for finance firms and fintechs to build higher-value AI services on top of it. This is an inference, but a strong one. (apicentre.paymentsnz.co.nz)

Trend 4: Governance is moving into operating structure

Westpac’s dedicated AI leadership role, FMA’s planned thematic review, and RBNZ’s continued focus on third-party dependency and cyber risk all point to AI governance becoming part of core operating architecture. (westpac.co.nz)

Overall Assessment

As of July 13, 2026, AI in New Zealand finance is best understood as operationally established in narrow but important domains, increasingly measurable in frontline service settings, and still bounded by trust, governance, and data infrastructure constraints. Since the June 10 edition, the biggest change is not a leap into autonomy but clearer evidence of organisational and ecosystem maturity: named executive AI ownership at Westpac, measured service gains at Tower, and more concrete open-banking readiness across the market. (westpac.co.nz)

The dominant New Zealand model remains AI as copilot, detector, summariser, workflow engine, and service assistant. The firms best positioned for the next phase are likely to be those that combine AI with secure consented data access, faster payment rails, strong fraud controls, and explicit human accountability, rather than those pursuing autonomy first. That conclusion is an inference from the current regulator, bank, insurer, and payments evidence base. (fma.govt.nz)