The glossy claims-tech pitches keep coming but the reality landing on broker desks can tell a different story. Insurers have bought the platforms, hired the data scientists and bolted AI onto everything that moves – yet the data can still be trapped behind departmental walls, complaint volumes are still smashing records and a new wave of deepfake-enabled fraud is forcing a rethink of what "evidence" even means. For brokers caught between clients demanding faster outcomes and carriers wrestling with their own internal plumbing, the message from the latest Claims Leaders Summit – and from the regulators watching it all – is that the industry's three biggest claims pain points are now colliding at once.
Start with the silo problem, which can be the single biggest reason promising technology is stopped in its tracks from delivering on its business case. The issue is not the tech – it is the organisational chart.
"Yes, definitely a big problem," said Nicola Brownlie (pictured), Nearmap’s senior solution engineer for insurance. "Even just the question of who we speak to, is a challenge because when you go beyond the 'pretty picture' or the data itself, the question is: how do we make sure we connect it to all the right teams internally and all the right use cases?"
It is a frustration echoing across the global market. Capgemini's World Property & Casualty Insurance Report 2026 found only 10% of P&C insurers are successfully scaling AI, and analysts point squarely at structural fragmentation rather than a lack of ambition. As one industry commentator put it bluntly, the industry "doesn't have an AI adoption problem—it has a structural design problem", with claims, underwriting and policy administration systems still speaking different languages.
Brownlie sees the consequences first-hand when Nearmap's aerial imagery and AI condition scores – data points that can confirm whether a shed was destroyed by a cyclone or already falling over, or whether a cash-settled customer actually fixed their roof – are loaded into a claims environment but never make it upstream to pricing and underwriting.
"One of the biggest challenges we have is that when we go and speak to the siloed teams, internally there's not really one overarching person or team responsible to think that way," said Brownlie. "There are a few insurers that are starting to have a more centralised team that will take the data in and start to work internally with the business to understand how it's used, but I think there's still so much to be uncovered."
That fragmentation has a price and is being totalled up in public. The Australian Financial Complaints Authority (AFCA) received a record 111,373 complaints in the 2025 calendar year, a 14% jump on 2024 and the highest annual volume in its history. General insurance complaints surged about 20% to roughly 36,000, with claim delays the number one issue across all financial services and denial-of-claim disputes accelerating fast, up 32% to 6,362. The first quarter of 2026 has already put the system on track for a record 120,000 complaints. afca.org.au
For brokers, the read-through is uncomfortable: clients are no longer quietly accepting decisions. They are testing wordings, evidence requirements and communications quality – exactly the areas where a broker's documentation discipline at placement and claims advocacy at the back end can make or break a renewal.
So how does the industry get the trend line bending the other way? According to one of the technology partners working behind the scenes for major third-party claims managers, the unglamorous answer is information hygiene.
"In terms of reducing the number of complaints within claims, it is largely about data continuity: getting more accurate information from the claimant from the beginning to the end," said Isaac Alexander, director of growth and innovation for April9, a Brisbane-based insurance focused software developer.
The logic is brutally simple: cleaner data in, faster decisions out, fewer disputes escalating to AFCA – and fewer awards of non-financial loss compensation, which Deborah Jenkins, AFCA's chief customer officer, recently warned land squarely on insurers and indirectly on the brokers who placed the cover.
If silos and complaints are today's problem, the third front is moving faster than either. AI is reshaping the rules of evidence and regulation in the same breath – and most carriers are still drafting their playbook.
"For me, it is how they are going to manage an ever-changing regulatory and technology space as AI continues to disrupt the world we live in," said Lincoln Grace, product manager from Wilbur, the claims management software provider. "I think that is going to have substantial effects on not only processes within insurance organizations, but also the regulatory changes that might come on the back of that."
A recent Verisk study found 99% of insurers have encountered manipulated or AI-altered documentation and 76% said altered claims submissions have become more sophisticated in the past year. Swiss Re's latest SONAR report warns that insurers are reporting rapidly rising use of deepfakes in claims fraud, skewed toward low-value claims – the very claims most likely to slip through automated workflows. Personal property lines look particularly exposed, with one Verisk director noting homeowners, renters and mobile home claims as the biggest opportunity for deepfake-driven fraud because there is usually a single party and no witnesses.
The takeaway from all three fronts is that the carriers that win the next claims cycle will be the ones that finally connect their data across silos, communicate transparently enough to keep clients out of AFCA's queue and treat AI as both a productivity lever and a fraud vector.
Nicola Brownlie, Isaac Alexander and Lincoln Grace were all interviewed by IB at this month's Claims Leaders Summit
