L to R: Luke Balmforth (Partners Life), Dan Freeman (Partners Life), Simon Baker (AIA), Adrianne Lowe (PPS Mutual), Kathie Tutty (Asteron Life).
On Thursday, 26 February, we gathered the insurtech community at Partners Life in Auckland to talk through a topic that’s very much front of mind right now: how we connect medical data and insurance in ways that are efficient, fair, and respectful of people’s privacy.
New Zealand has seen two significant health data breaches in recent months, and new privacy regulations are coming into force in May 2026. Against that backdrop, the room was full of people who work with these issues every day — underwriters, data specialists, transformation leads, and insurtech founders.
Joining the panel were Dan Freeman, Head of Underwriting Service at Partners Life; Adrianne Lowe, Chief Underwriter at PPS Mutual; Kathie Tutty, Chief Underwriter at Asteron Life; and Simon Baker, Head of Transformation at AIA. The discussion was moderated by Luke Balmforth, Chief of Staff at Partners Life.
What Does “Proportionate” Data Actually Mean?
The panel opened with a question that sounds simple but isn’t: what counts as proportionate health information when assessing risk, and why does the answer vary so much between insurers, and between insurers and GPs?
On one level, everyone agrees on the principle. Insurers should only request what’s genuinely necessary. But in practice, underwriters are managing uncertainty, and when in doubt, the instinct is to ask for more rather than less. Add in differences in risk appetite and reinsurance requirements, and you can have two insurers with similar profiles requesting quite different levels of information.
The relationship between insurers and GPs adds another layer of complexity. GPs write notes for clinical care, not for risk assessment. So underwriters might receive pages of records that include personal details or an unrelated minor injury that have no bearing on the cover being applied for. One panellist summed it up neatly; the two professions are using completely different lenses, and that gap creates friction for everyone, including the customer.
The Data Quality Problem
One thread ran through almost every part of the discussion: getting access to medical data is only half the challenge. The quality and structure of that data matters just as much.
What underwriters actually need is fairly specific: a diagnosis, a date, a prognosis, and an understanding of functional impact. What they often get is dozens of pages of unstructured clinical notes written for a different purpose entirely. If the industry could agree on cleaner, more structured data formats, it would mean faster decisions, more consistency, and a better experience for customers waiting to find out whether they’re covered.
AI came up naturally here. The panel saw genuine potential in using AI for data extraction and summarisation — pulling out the relevant details quickly so underwriters can focus on the decision itself rather than the document triage. But there was also clear-eyed caution. In areas like income protection and mental health underwriting, the judgment involved is nuanced and contextual in ways that current AI models aren’t well-suited to replicate. A transparent, auditable process matters to customers, advisors, and regulators alike. A system that produces outcomes nobody can explain is unlikely to hold up.
AI governance was also raised as a practical concern. When large language models are freely available to staff, you need clear policies about what can and can’t be done with sensitive medical information. And those policies need to keep pace with tools that are evolving quickly.
Privacy as an Opportunity, Not a Constraint
The upcoming IPP3A amendment was covered too. From 1 May 2026, organisations will need to notify individuals when their personal information is being collected from a third party. For insurers, this means consent forms will need to be more explicit: clearly stating the purpose, naming who the data will be shared with, and separating consent for underwriting from consent for claims.
Rather than treating this as a compliance headache, the panel’s view was that it’s worth approaching as an opportunity. Customers are paying closer attention to what happens to their personal information than they were five years ago. Insurers that take privacy seriously and demonstrate that clearly are in a better position to earn lasting trust.
Where Insurtechs Come In
Unstructured data, inconsistent information requests, evolving consent requirements, and AI governance are challenges that the right technology partners can help address.
There’s meaningful space for insurtechs to work alongside insurers on tools for structured medical data exchange, smarter consent management, and AI-assisted underwriting that keeps the human in the loop. The panel made it clear that insurers aren’t looking for off-the-shelf fixes. They’re looking for partners who understand the regulatory environment and the operational realities, and can help move things forward in ways that actually work.
Get Involved
Thank you to Partners Life for hosting the event. If the themes from this event resonate with you, we’d love to hear your perspective. InsurTechNZ is a community built on shared learning and honest conversation, and it’s stronger when more voices are part of it. Join InsurTechNZ and help shape what innovation looks like in New Zealand insurance. Find out more at insurtechnz.org.nz.
A recording of the panel discussion is available for members of InsurTechNZ.
Next InsurTechNZ Connect event: 23 April, hosted by Deloitte — check the website for details.
