Generative AI and its impact on the claims handling sector

"I do think we're at a similar inflexion point to the dawn of the internet"

Generative AI and its impact on the claims handling sector


By Mia Wallace

In preparation for his discussion with Insurance Business on the impact of generative AI on insurance claims handling, Benedict Burke (pictured), chief client officer, global client development, at Crawford & Company went straight to the source. Loading up his OpenAI account, he asked ChatGPT that same question – and in 90 seconds received a lengthy reply, “not one word of which I disagreed with.”

This is a real example of the power of the tool in action, he said, and it provides an excellent springboard for a meaningful discussion around the threats and opportunities represented by generative AI.

Artificial intelligence – reshaping clients’ expectations

In his role, Burke supports a diverse range of global businesses operating in a myriad of sectors from leisure, to pharmaceuticals, to heavy manufacturing, and the disparate and rapidly evolving nature of these clients’ dynamics demands that Crawford keep pace. What will be interesting to see, he said, is where AI will fit in with that requirement – and how it might reshape clients’ expectations.

“So, is AI a threat or an opportunity? Without a doubt, and I think this is a widely shared opinion - it offers a huge opportunity,” he said. “I think in many ways, this is analogous to the birth of the internet, we are at an inflexion point. My grandson, who is less than a year old, will find his whole world governed by generative AI and machine learning, and his lifestyle will be fundamentally based on this enormously powerful technology.”

With that in mind, Burke believes that the insurance sector – and especially the insurance claims sector – should be looking at the most effective ways to harness these capabilities. Critical to note, he said, is that the opportunity of generative AI is not limited to just creating efficiencies but also enabling consistencies in the claims handling process that are to the benefit of the policyholder.  

Used correctly, he said, it can become a “digital assistant”, enabling loss adjusters and claims handlers to gather, correlate and understand huge amounts of data in an instant – and to use that to generate outcomes that will ultimately improve the customer journey. A look at some of the key metrics around trust in insurance quickly reveals why this customer-centricity piece is so crucial.

“Only 55% of policyholders think their claim is going to be paid at the outset. That’s the reality of where trust in insurance stands right now,” he said. “And we’ve got to tackle that perspective because we’re seeing growing expectations from consumers about how the products they’re buying should be provided to them in terms of efficacy and delivery. Expectations around 24/7 capability, transparency, openness and access are growing, while trust is declining.”

The rapid advancement of automation offers a unique opportunity for insurance businesses to plug that 45% gap in consumer trust, he said. And it’s thrown into further relief by the metric that 80% of people who have had a poor experience of claims handling will move their insurer the following year.

How Crawford is utilsing generative AI

Burke highlighted three key strategic pillars that form the foundation of Crawford’s operating model – championing customer-centricity, setting industry benchmarks in terms of quality, and driving real expertise into and through its business.

Ultimately, AI will have an impact on every part of that strategy – enabling greater efficiency and consistency in how Crawford delivers a service that consumers trust and which is industry-leading in terms of quality, without losing the unique proposition of its expertise.

“In our network solutions business in the US, where we’re piloting a lot of this AI technology right now, it’s digesting a huge amount of data at the first notification of loss (FNoL) stage and now triaging those claims automatically,” he said. “And that means we can immediately route that claim to the right skillsets and into the right processes which ultimately will provide the best outcome for the customer and our clients.”

Fundamentally, Crawford’s clients want two things from the business, he said, they want Crawford to treat their customers fairly and they want to pay a fair amount under the policy wording. Implemented correctly, this technology will allow companies to get ahead of claims trajectories which at aggregate or portfolio level are going in the wrong direction in terms of costs. And it will do so by allowing Crawford to understand the root causes of that – whether it’s a lack of alignment, the wrong processes being implemented, or external pressures such as inflation.

“Essentially, this will lead to a fairer outcome for the policyholder and to insurers recognising that Crawford can demonstrate that our processes and inputs are adding true value both at the customer level and in the context of reserve management, which is critical for insurers,” he said. “So, having a more accurate and consistent understanding of what’s happening at a claims portfolio level will be a huge benefit that derives from AI.”

Challenges around generative AI

However, the opportunities which abound do not erase the case for caution when it comes to grappling with the full implications of AI, he said. Machine-learning will fundamentally shift existing operational models and processes, and because the tool is self-learning, care must be taken to ensure the machine doesn’t embed prejudices or inaccuracies into its modelling structures. Other areas such as data security also highlight the need for effective regulations and controls.  

Burke emphasised the need for sector-specific controls and regulations around generative AI as the nuances required for each industry sector will differ so significantly. Essentially, he said, the industry works better with regulation and it’s clear that regulation is going to have to play catch-up very quickly in order to keep pace with the rate of evolution happening within generative AI.

The impact of AI on claims handlers and loss-adjusters

Where AI represents both a challenge and an opportunity, is around the impact it will have on the people who make up the claims space. At Crawford, senior managers have three principal duties, he said, to look after the interests of their customers, their shareholders and their people.

For Burke, this final piece is the most critical and he’s especially interested in how AI will change the workload and lifestyle of Crawford’s internal teams. Assessing the capabilities of AI, he said, it’s clear that in the next 10-15 years, it could lead to some disintermediation of claims handlers’ roles, particularly in the context of low-complexity claims. However, he strongly affirmed the essential role a human touch will continue to play in the handling of a claim.

Empathy, understanding, nuance, negotiation and project management skills are all components of an experienced and skilled adjuster, he said, and the value-add of those skills cannot be replaced. Rather, AI will take up its natural place as a digital assistant to the loss adjuster of the future, and take away some of the “heavy lifting” of the administrative burden which currently takes up time these skilled individuals would rather be using settling claims and ensuring the efficacy of the products being sold.

“I think AI will become our author, but our loss adjusters will remain the editor. The AI will present data and options which the adjuster can interpret and add value to,” he said. “And we talk a lot about technology and people combining in Crawford to produce an environment where we’re best-in-class. And which adjusters wouldn’t want to be in an environment where they get to use their intellect, skills, experience and negotiation capabilities, and not have to deal with the mundanity of data analysis?

“I do think we’re at a similar inflexion point to the dawn of the internet. It’s hugely positive, it’s hugely exciting and I’m keen to see how it all comes through. I think it will take five-to-10 years to start to really understand that. I see it being very quick, radical and in some respects dramatic at first. But then there’ll be a period of settling down and regulation where it comes more normalised and institutionalised. It’s certainly a highly exciting time to be alive.”

What are your thoughts on the implications of generative AI for the insurance sector? Feel free to share them in the comment box below.

Related Stories

Keep up with the latest news and events

Join our mailing list, it’s free!