Claims was traditionally a slow-moving source of insight in insurance: messy narratives, inconsistent documentation and lessons that arrive months later - after the loss ratios have already shifted and renewal terms have hardened. But AI is compressing that cycle. For insurers, it promises earlier warning of emerging patterns and faster operational pivots. For brokers, it raises a sharper question: if the market can see problems forming sooner, what does “good advice” look like before the losses land?
AI’s growing influence across the entire claims lifecycle will be a major focus at the Claims Leaders Summit in Sydney on May 12, where speakers and panels are set to examine how AI is reshaping claims response, fraud detection and the regulatory expectations developing around its use.
Eric Lowenstein (pictured), CEO of Sydney-based healthcare and medical indemnity underwriting agency Tego, said the pace of change triggered by the technology in recent months has become a story in its own right for the insurance industry.
“The pace of AI and technology is moving faster than any of us can comprehend,” he said.
That acceleration is now reshaping claims workflows and the way underwriting learns from them - particularly in portfolios where small shifts in frequency, severity or litigation behaviour can quickly change pricing and appetite.
Insurers are increasingly positioning AI as a way to move claims intelligence closer to “live” - not necessarily repricing risks hour-by-hour but identifying loss signals earlier and feeding them back into triage, reserving, loss control, underwriting rules and broker conversations.
That direction of travel is visible in claims operations research and industry commentary.
In Unveiling The Carrier Perspective: 2025 Claims Insights Whitepaper, Gallagher Bassett highlighted generative AI as a growing feature of claims management, pointing to expanding adoption and day-to-day use cases across the claims lifecycle. Even where AI begins as summarisation, document handling, or triage, the practical outcome is the same: cleaner data, faster classification of causes and contributing factors and earlier detection of recurring patterns.
For brokers, the significance is less about the software and more about timing. Faster feedback loops can change when insurers intervene (regular risk advice rather than just at renewal), how quickly exclusions or endorsements appear after a new claim trend and how quickly premium pressure becomes visible in targeted segments. It can also reshape “expectations management” after an event -when customers assume recovery will be straightforward but the claim environment through factors like repair capacity, fraud spikes, litigation or inflation is deteriorating in the background.
At the market infrastructure level, Australia is also moving toward more coordinated, real-time intelligence sharing in at least one high-impact area: fraud. In a November 2025 news release, the Insurance Council of Australia (ICA) announced a collaboration with Shift Technology and EXL to build a national data analytics fraud detection and investigations platform, starting with motor claims and designed to alert investigators to suspicious activity “in real time.” If that capability scales, it is effectively a faster feedback loop for the entire industry that could influence claims handling posture and, eventually, underwriting settings.
As AI-driven signals arrive faster, brokers may find themselves pulled into more frequent, more specific risk conversations. If an insurer’s claims data shows a spike in a particular loss driver - certain injury types, contractor failures, or a fraud pattern in a region - brokers may be asked to support mid-term risk controls, updated declarations, or sharper client communications. That can help protect clients, but it also raises the bar: earlier warnings can reduce the tolerance for “we didn’t know.”
The same shift increases the importance of trust and data discipline.
“Just sending your data into an LLM model is a risk,” said Lowenstein.
It's a reminder to the industry that AI use in claims and underwriting can only be scaled safely if insurers and their distribution partners, including brokers, can demonstrate governance over privacy, data handling, model oversight and human review.
New Zealand’s regulators are signalling a similar stance: enable innovation but demand oversight. The Reserve Bank of New Zealand (RBNZ) has said AI adoption is accelerating and that AI can improve risk assessment and cyber resilience, while also posing risks including AI-driven errors, privacy concerns and concentration risk from reliance on a few third-party providers. The Financial Markets Authority (FMA) has also emphasised governance and oversight, encouraging firms to ensure AI outputs remain reliable, explainable and contestable, and that sensitive information is protected.
Meanwhile, major brokers are productising the “insight-to-action” promise. In November, Aon, for example, announced an AI-enabled claims platform positioned around claims resolution analytics and consistency - an indicator of where broker value propositions are heading as clients demand faster clarity and outcomes.
Lowenstein says implementation is already moving from theory to tooling: “We have recently launched our own claims platform with Five Sigma, which has quite a lot of AI elements to it, as well as a new underwriting platform.” For brokers, the takeaway is that AI-driven claims intelligence is no longer a future capability; it’s becoming embedded in how decisions are made.
The near-term opportunity is faster detection, earlier intervention, better outcomes. But the strategic shift is sharper: in a market where signals arrive sooner, insurers and brokers alike may be judged on how quickly - and how responsibly - they act.
