New Zealand banks are replacing 1960s-era core systems while staying online. Image / 123rf
KPMG technology consulting partner Rachael Niao compares upgrading a bank’s core system to heart surgery: “It is really, really tricky … you’re tearing out the heart of your bank."
She says all the major New Zealand banks are in the process of upgrading core systems. “They are making deep structural changes to those systems and modernising them to unlock innovation.”
It is the first time so many of New Zealand’s major banks have attempted this scale of change and, unusually, they are all going through transformations at roughly the same time.
“It’s not just about creating agility and resilience, or to be digital first, but it’s also about being able to use technologies like AI.”
The upgrades are a huge challenge for banks given how old their systems are, in some cases there are technology components dating back to the 1960s. Yet Niao says there is also a massive opportunity.
Niao’s experience gives her a distinctive perspective into the capital intensive, multi-year projects that are underway. She worked overseas on major banking transformations for Accenture in Europe, Asia and Australia. Then worked for EY in Australia before returning to New Zealand four and a half years ago.
During her time in Australia she worked on multi-billion banking transformation programmes for two banks. “The initiatives aimed at modernising legacy systems so they could enable the digital banking that you see there.”
This was about a decade ago. She says the customer experience, the bank apps and the way banks bring products to market is more advanced in Australia than what you see here.
It wasn’t always that way. In the 1990s, New Zealand was a global leader in banking innovation. We were among the first countries to adopt Eftpos and pioneered the move toward cashless transactions.
“Back then, we were the benchmark for electronic transactions,” says Niao, who recalls using her Eftpos card routinely in New Zealand before moving to the UK, where cash still dominated. “We seem to have been leapfrogged.”
The core banking systems many New Zealand banks operate are, by technology standards, ancient. Some still run on Cobol, a programming language so old that only programmers in their 60s and 70s retain the skills to support it. Nobody young wants to learn it, “those systems need to go”, says Niao.
Legacy platforms mean even small changes are expensive and slow. They weren’t designed for today’s digital ecosystem, struggle with modern compliance requirements and are a barrier to the adoption of emerging technologies like real-time payments and AI-driven services.
The gap between Australia and New Zealand bank systems exists partly because of ownership structures. While Australian parent banks own most major New Zealand banks, they acquired rather than built them.
“The systems are quite different. The Australian banks bought our banks. They didn’t co-create them, and so they are on very different technology stacks.” Australian banks completed their transformations about a decade ago but are only now rolling similar changes out to their New Zealand subsidiaries, which still need to navigate local regulatory requirements and legacy constraints.
Getting it wrong can be catastrophic. Niao points to a UK bank which experienced “everybody’s worst nightmare”. When its core system migration failed, customers couldn’t access accounts for months. “Once you go live, you can’t go back. You have to fix forward. It pretty much broke the bank.”
The arrival of practical AI technology has ramped up the urgency for transformation. Unlike previous technology waves, AI’s power depends entirely on the underlying infrastructure.
The arrival of practical AI technology has intensified the urgency for transformation. Unlike previous technology waves, AI’s power depends entirely on the underlying infrastructure.
“AI is only as good as the data you have,” says Niao. Legacy systems store data in formats that make it difficult for AI algorithms to access and process efficiently.
Modern cloud-enabled platforms, by contrast, allow AI to run sophisticated algorithms across massive datasets in real-time. Older systems can’t cope.
This is why Australian banks, having modernised their core systems, can now deploy AI at scale while New Zealand banks remain limited to pilot projects and isolated use cases.
Niao says that means they can use AI for credit risk modelling, liquidity forecasting and customer engagement.
New Zealand banks are beginning to follow suit within their constraints. ASB uses AI for real-time fraud detection. Westpac employs AI algorithms to streamline regulatory compliance reporting, reducing both operational risk and the workload for staff dealing with increasingly complex requirements.
But these remain “pockets of AI” rather than full adoption. “They’re still not able to leverage it across everything, which you see in the Australian market,” says Niao.
Once transformations are complete, the opportunities expand to hyper-personalisation of products, AI-powered support and predictive analytics for everything from credit assessment to detecting cyber threats.
The good news is that New Zealand banks’ smaller size could be an advantage. “Because they’re smaller and more agile than their Australian counterparts, that agility gives them an opportunity to leapfrog and adopt AI solutions much faster,” she says.
Niao says human oversight remains essential. Like many of us, she now uses AI daily in her consulting work to pull information together quickly. It has dramatically improved her productivity. “But I don’t just send it out. I always check. I always ask it to cite sources and then verify those sources are right.”
The Reserve Bank and Financial Markets Authority are establishing rules around AI use in banking. These emphasise transparency and risk management. Customers expect fairness and clear explanations for AI-driven decisions. “There still needs to be human intervention,” says Niao, “because AI hallucinates and doesn’t always get things right.”
While transformation on the scale banks operate at is both risky and expensive, there is a proven pay off. Niao worked on the Commonwealth Bank of Australia’s (CBA) core banking transformation from 2008 to 2012. Starting during the Global Financial Crisis, it cost far more than expected and, at the time, drew public criticism.
“Yet when you look at where they are in the market compared to all the other banks, the return on that investment has been tenfold,” she says. CBA has been Australia’s No 1 bank ever since, with no regrets.
For New Zealand’s banks, the next few years could determine whether they reclaim the innovation leadership they once held. “Once the constraints are gone,” says Niao, “the opportunities are massive.”
While New Zealand banks grapple with AI adoption, Māori communities are advocating for a fundamentally different approach to data governance. It treats data as taonga, an extension of a person’s wairua or spirit.
“It’s about how you store that data, how you use the data, how you protect the data and how you dispose of that data,” says Niao, who is deeply involved in Māori data sovereignty issues.
She gives an example of some government agencies that now perform karakia over data belonging to deceased individuals before disposal, acknowledging the person and their passing.
This approach contrasts sharply with the transactional model of big tech companies. Māori data sovereignty principles emphasise co-design with iwi and hapū, embedding tikanga into data governance frameworks to ensure equitable outcomes.
The principles are particularly important for government agencies holding sensitive information about Māori in justice, health and social services. Niao says: “We just want that data treated like a taonga”.
KPMG is an advertising sponsor of the Herald’s Business Reports.