Spotlight: Unstable and rising claims drive insurer automation

As inflation, fraud and extreme weather push up claims’ volumes and costs, insurers are turning to automation and AI to manage demand, improve service and stay competitive, writes Chris Marshall.

In many ways, 2024 was a record-breaking year for the insurance sector – though not in ways anyone would have hoped.

Vehicle repair costs reached a new high of £7.7bn and car insurance claim pay-outs hit a record £11.7bn. Meanwhile, amid the highest number of flood warnings ever recorded, payouts for weather-related damage to homes and possessions soared to an unprecedented £585m.

Claims inflation – driven by factors including climate volatility and supply chain disruption – shows no sign of slowing. “We see an increase in the severity and frequency of claims based on the weather and climatic events,” says Rémy Gounel, subject matter expert, claims at Shift Technology. “We see inflation in costs for property repair claims, and for motor, where there is increasing use of technology,” he says. 

Caught between rising claim volumes and intensifying cost pressures, insurers are adapting. They are working to capture the benefits of technologies such as machine learning, predictive analytics, and generative AI (GenAI) to automate, streamline and improve the claims process. But it is not always easy. 

This approach enhances customer satisfaction by expediting straightforward claims and also ensures that intricate cases are given the time they require.

Karl Parr, claims technical and central services director, AXA UK

The case for claims automation

Manual processing of large volumes of unstructured insurance claims data, from various sources, is time-consuming and repetitive. Claims handlers spend about 30% of their time on low-value work, such as reviewing documents, according to Shift Technology.

When the number of claims increases – say in the aftermath of a major flood – then inevitably teams struggle to process a high volume of claims quickly enough, and customer service can suffer.

So now, many insurance companies are working hard to automate routine tasks, streamline their claims management process and free up their agents to concentrate on building stronger customer relationships.

Take the example of Axa UK, which has automated its claims assessment system. Karl Parr, claims technical and central services director at the insurer, says technology offers a “huge opportunity” both to improve service and to help colleagues focus on more complex tasks. He describes it as a “pivotal strategy to manage claims volume and reduce costs effectively”.

“This approach enhances customer satisfaction by expediting straightforward claims and also ensures that intricate cases are given the time they require,” says Parr.

Axa UK continues to explore new ways to automate and use AI. One, Parr says, is in predictive analysis, “helping us better anticipate claim trends and adjust our resources accordingly”.

Human meets machine

Just like companies in many other sectors, insurers on the road to automation tread a thin line between becoming more cost- and time-efficient and keeping the human touch where it is needed.

“Humans remain irreplaceable in our operations,” says Parr. Customers can speak to a human at any point they want to. He argues that the complexities of individual claims often require “empathy, understanding and nuanced decision-making” – something that a robot can’t provide.

Of course, Axa UK isn’t alone in striving to balance these competing forces. Richard Napoli, claims and legal services director at business insurer Markel, lays bare the choices facing insurers: “Without considering automation in the claims process, the only way to manage increasing volumes of claims is to increase resourcing to match the workload or lower your service levels. Either option risks the success of the insurer as a forever growing headcount is unsustainable and lowering standards is damaging to reputation.”

For this reason, Napoli agrees high-volume insurers are right to consider automating high-volume claims that can be filtered into a standardised process. But he says that approach isn’t right for all insurers. 

At Markel, which usually deals with one-off, complex claims, “our use of technology is primarily related to back-office systems and processes helping us be more efficient”, says Napoli. “We don’t shy away from using technology in the right circumstances, but we very much see it as an enabler.” He adds a word of caution: “Automation and AI add little value in complex, non-binary, people-centric claims.”

Yet, research suggests not all insurers see a need for human involvement in automation, at least not in some areas. According to a 2024 survey by RDT, only 40% of respondents considered having a ‘human in the loop’ as business-critical for automated workflows and 15% deemed this to be unimportant.

And what about policyholders – how much do they care? More than half (59%) of consumers, responding to a survey (specifically related to health insurance) said they would be open to an AI-driven claims process if it meant faster, more accurate resolutions. While that suggests a majority of people may be okay with automated processes, that still leaves many doubters.

Some uses of automation are perhaps less controversial. This includes accessing real-time supply chain data during surge events. More than half (54%) of insurers responding to a 2024 survey for Gallagher Bassett had implemented digital claims processing to address supply chain challenges.

We don’t shy away from using technology in the right circumstances, but we very much see it as an enabler.

Richard Napoli, claims and legal services director at Markel

Smarter fraud detection

As the volume of claims across several different insurance lines rises, so does the risk of fraudulent claims. To take one example, Aviva reported a 14% increase in the number of claims it declined due to fraud in 2024, uncovering more than 12,700 suspect claims worth £127m.

“Insurers have to acknowledge that this kind of unpredictability correlates with increased fraud exposure,” says Gounel. “People know the process is simplified in those moments. With Shift, insurers can quickly detect new fraud indicators, even if they’re unique to an insurer’s book of business and geographic footprint.”

Opportunistic fraud often spikes in the aftermath of weather catastrophes, when claims volumes surge and information may be incomplete or ambiguous. In such conditions, traditional responses can fall short – leading to delays for genuine claimants, overburdened contractor networks, and slower community recovery.

While automation is helping streamline many claims processes, it also presents new challenges, as fraudsters adopt increasingly sophisticated techniques to falsify evidence – from digitally manipulated images to doctored documents. That’s why adaptive, intelligent fraud detection has become a critical part of modern claims handling. 

Rory Yates, global strategy lead at EIS, gives the example of an insurer that standardised fraud detection across more than 2,000 claims experts, and reduced fraud costs by over 40%.

Old systems, new problems

Many insurers are beginning to reap the benefits of automation in their claims processes and beyond. But, as shown by the challenge of balancing human and AI inputs, it is not easy and will not happen overnight. The industry is still dipping its toes into the potential offered by AI and automation.

One of the common hurdles to quicker uptake remains how to make automation work with legacy technology. According to Earnix’s 2024 Industry Trends Report, modernising legacy technology systems remains the biggest challenge for insurers.

“It is hard to plug new generative AI or rules-based modules into legacy systems,” says Shift Technology’s Gounel. “It requires a lot of money and time for IT teams if they want to do it internally. Intelligent AI platforms that are capable of offloading a significant part of the analysis and decision process into an integrated SaaS layer – but still flexibly communicate back to the decision process in the legacy platform – provide a mechanism to overcome this.”

At the beginning it was difficult, but now more and more insurers realise automation is not there to replace people but to supercharge them.

Rémy Gounel, subject matter expert, claims at Shift Technology

What of the fear that AI and automated processes will take jobs? Gounel believes attitudes among workers in the sector are changing: “At the beginning it was difficult, but now more and more insurers realise automation is not there to replace people but to supercharge them.”

In a world of rising claims and relentless change, some supercharging is exactly what insurers need.

Sponsored content