Speech by Kari Jones to KiwiSaver Providers, AI and the Future of Financial Services

Kari Jones, Executive Director, Financial Markets Authority (FMA)

Introduction

Tēnā koutou katoa and thank you for the warm invitation to speak tonight.

I want to begin by acknowledging the AI Forum our hosts this evening. I’ve been involved in this community for many years, including as a former Executive Committee member, and I know first-hand how important your work is. I especially want to recognise your contribution to shaping the Government’s emerging AI Strategy alongside Minister Collins.

As many of you know, I’ve been speaking publicly about AI for some time now  and I remain a bullish proponent of the transformative role it can play across our economy, our public services, and our everyday lives. To that end, the AI ecosystem offers present and emerging opportunities for long-term value creation for NZ business and consumers alike.

Alongside my role at the FMA, I also serve on the board of TUANZ. In our Digital Priorities 2025 report, we found that New Zealand businesses including financial services are slow to adopt emerging technologies like AI. That report called for:

• Accelerated technology adoption
• Effective harnessing of data and AI
• Improved digital infrastructure and skills development
• And most importantly: stronger partnerships between government and industry

As far back as 2018, I gave the keynote at the Corinium Data & Analytics Conference urging New Zealand to embrace AI. Today, that call feels more urgent than ever.

So why is someone from the FMA talking about AI?

While I’ve only been at the FMA a few months, I’m already seeing the immense opportunity AI presents not just for the firms we regulate, but for regulators themselves. Our role is to strike the balance: enabling innovation that improves outcomes for customers while maintaining trust, transparency, and integrity in the system.

And AI sits right at that intersection.

Used well, it can lift financial literacy, lower costs, and personalise services for consumers. But it must be deployed responsibly, with appropriate governance, risk management, controls and policies – placing customer outcomes front and centre.

What’s Happening Offshore

Let’s start globally.

Other jurisdictions are moving fast. According to the Bank of England, many financial services firms are already using AI to automate internal operations, such as code generation, document processing, and customer service. But more significantly, they are beginning to integrate AI into core financial decision-making, including credit underwriting, pricing, and capital allocation.

This points to a strategic shift, AI is no longer just an efficiency play; it's increasingly becoming a driver of strategy for financial services.

In the UK, the Financial Conduct Authority (FCA) recently launched AI Live Testing, allowing firms to test generative AI models in collaboration with supervisors. The goal is to build confidence in AI tools while exploring new evaluation methods that support responsible, consumer-facing deployment.

We at the FMA share that ambition.

Our own regulatory sandbox recently launched with six firms selected. AI was of the top three themes that cut across all of the applications in this first round, it’s only a matter of time before AI-powered fintechs dominate the next cohort.

The regulatory sandbox approach will help us to better understand the opportunities as well as barriers to innovation for firms across the AI value chain – from development to release and deployment.

AI in Fund Management and KiwiSaver

So what does this mean for you, as KiwiSaver providers?

Globally, we’re seeing AI reshape investment operations from enhanced valuations of private assets to portfolio optimisation and real-time risk management.

As KiwiSaver funds increase exposure to private markets, AI tools offer a promising way to improve valuation accuracy, transparency, and due diligence, while enabling more sophisticated and faster investment decisions.

Take Vanguard Digital Adviser as an example. It uses AI to provide tailored advice based on life stages, ethical preferences, and retirement goals. This promotes accessibility to advice and enables tailored guidance for members with low balances that are traditionally a segment underserved by the market.

I see this as the start of a major opportunity for KiwiSaver.

AI can:

• Attract new members through personalised marketing
• Provide smarter onboarding and product guidance
• Offer 24/7 support via chatbots and digital assistants
• Enable hyper-personalised retirement pathways

The ability to lower the cost of service while lifting the quality of engagement is no longer theoretical it’s here.

And while these advances offer commercial value, we expect providers to reinvest efficiency gains into improved customer experiences especially for those currently underserved.

Tackling Structural Inequities in KiwiSaver

Let’s talk about equity.

Recent Retirement Commission research revealed a persistent 25% gender gap in KiwiSaver balances up from 20% in 2022. More women than men have balances under $10,000, and more men than women hold balances over $80,000 across nearly all age groups.

This is a systemic issue but AI may offer part of the solution.

Imagine a KiwiSaver experience that:

• Sends personalised nudges before and after parental leave
• Provides contribution planning for those re-entering the workforce
• Offers flexible, gender-sensitive calculators and projections

When we personalise services through AI, we create the opportunity to open the door to fairer, more inclusive financial outcomes.

And it’s not just hypothetical.

A recent Financial Planning Standards Board global survey of more than 6,200 financial advisers across 24 countries (including NZ) found:

• 60% expect AI to enhance the quality of advice
• 80% say it will help them better serve clients
• 60% believe it will expand access to financial planning for underserved populations

This is where AI becomes a true social leveller and as regulators, we welcome it.

The FMA’s Use of AI

We’re not just encouraging others to use AI. We’re also embedding it into our own operations.

At the FMA, we’ve begun:

• Using AI to scrape and analyse data for a new MIS entity risk model
• Analysing feedback themes from industry engagements in our CRM system
• Exploring how AI might assess and summarise marketing and disclosure materials for consistency and compliance

Our aim is to become a more targeted and efficient regulator focusing effort where risk is highest, and creating space for low-risk innovation to thrive.

This isn’t about automating judgment. It’s about augmenting human capability so we can work smarter, not just harder.

Governance, Risk, and Responsible Use

Of course, with any powerful technology especially one as fast-moving as AI we must talk about governance and risk.

The good news is, this isn’t uncharted territory.

Many of you have already built strong governance processes for machine learning models especially in areas like “next best action”, fraud detection, and risk modelling. These frameworks include testing for bias, explainability, performance monitoring, and human oversight.

But generative AI is different. It produces content, not just outputs. It learns from broader and less structured datasets. And it can generate entirely new responses, rather than making predictions based on defined variables.

This introduces new questions:

• How do you govern probabilistic outputs?
• How do you audit AI-generated language or imagery?
• How do you explain decisions when the logic isn’t always linear?

We now have the opportunity and the obligation to revisit and upgrade those existing governance frameworks, ensuring they are fit for purpose in a world where AI powers personalised conversations, investment recommendations, or disclosure narratives.

That means:

• Clear ownership and accountability for AI models
• Cultural competence, particularly around the safe and respectful use of Māori language and perspectives
• Technical competence to assess design, limitations, and risks
• Transparency for customers when AI influences a recommendation or decision

Good AI governance requires more than compliance it demands ethical foresight, interdisciplinary skill, and a commitment to fairness.

It’s not a reason to hold back. It’s a reason to move forward with purpose and with care.

Conclusion and Call to Action

AI has immense potential to transform New Zealand’s financial services including making the KiwiSaver regime more inclusive, efficient, and member-focused.

At the FMA, we are:

• Laying the groundwork for safe and smart adoption
• Building internal capability
• And open to new ideas and partnerships

So my ask of you tonight is simple:

Talk to us.

• Share what you’re experimenting with.
• Tell us what’s working and what isn’t.
• Help us understand how we can best support responsible innovation.

Let’s ensure AI adoption in New Zealand delivers better outcomes for customers, not just better margins for firms.

It’s an exciting time to be working at the intersection of technology, finance, and regulation. And I’m delighted to say that, so far, my bullish enthusiasm is being fully vindicated.

Ngā mihi nui  and thank you again for the opportunity to be here tonight.