AI's impact on claims processes and procedures – Part 1

Editor's Note: This is another article in a multi-part series that examines the use of AI in the claims space. This is part one of a two-part series.

Takeaways:

• AI is expediting aspects of the claims process for carriers and customers.
• A key aspect of adoption is maintaining a human in the loop to review processes and decisions.
• Data accuracy plays a key role in the outcomes AI generates. 

Whether a company has fully embraced the adoption of artificial intelligence (AI) into its processes or is just gingerly entering into this space, the reality is that it is permeating every aspect of insurance. For claims, it has tremendous potential to expedite everything from the first notice of loss (FNOL), to reviewing and capturing the scope of the loss, setting the reserves and determining the final payout for policyholders. 

A recent study from Digital Insurance found that throughout the insurance industry, AI adoption is increasing, with more than half of the professionals anticipating at least a small-scale implementation of AI in the next 12 to 18 months across their companies. Their primary goal in adopting AI tools is to advance their operational efficiency and automation. Improving the customer experience also continues to be a major priority.

However, this transformation does have some challenges, such as a positive return on investment (ROI) for the implementation costs. The study also found that insurance organizations are more likely to see a positive return on investment from AI as compared to other financial services verticals such as the banking, mortgage and accounting industries.

Incorporating AI into property claims

According to the Verisk Quarterly Property Report, in the first quarter of 2025, claims volume hit a five-year low, with total claims dropping 7%, despite several significant weather events and the wildfires in California. However, replacement cost value increased 46% compared to the same period in 2024, due in large part to the Eaton and Palisades wildfires, which produced almost 48,000 claims at an estimated cost of $10 billion.

Following events like the wildfires in California, and major hurricanes, flooding or tornadoes that affect large areas, the use of AI both by insurers and policyholders plays a more important role in the claims process.

"AI has the ability to reshape the claims process from end to end—enabling insurers to move faster, work smarter, and deliver a better policyholder experience," shared Peter Miller, CPCU, president and CEO of The Institutes in an email to Digital Insurance.  "And as risk management and insurance shift toward a predict and prevent model, by leveraging technology to both predict and help prevent losses from occurring, AI plays an increasingly central role in both resolving and avoiding losses. However, AI can both be a tool in the fight against claims fraud, as well as a tool that further enables fraud to perpetuate."

Within the claims process, Miller says carriers are using AI in multiple areas such as for the FNOL where chatbots and digital intake tools can guide customers through the structured questions, help them validate their policy information, and capture supporting documentation such as photos or videos. AI can also help with prioritizing incoming claims and communicating with policyholders, but Miller stresses that human oversight is still required in all of these instances.

Pam Urueta, general counsel and chief claims officer at AXA XL, explains that insurers are increasingly embedding AI into their claims processes in an effort to drive efficiency and improve the outcomes for both the companies and their customers. "In commercial lines, the first places to start are intake and triage, where we can leverage AI to help extract data from documents to accelerate the claim creation process," she says in a statement to Digital Insurance. "Generative AI is being used by claim handlers to summarize lengthy reports, compare documents, and provide initial drafts of communications. We're also continually evaluating opportunities to leverage predictive analytics in areas like severity, segmentation, complexity and fraud."

The benefits of implementing AI into the claims cycle holds promise for both insurance claims professionals and their customers. "It's about how we can deploy AI into the claims lifecycle that has a dual benefit to our people and our customers," details Lee Elliston, chief operations officer/claims & UK operations for Aspen Insurance, in an email. "But we are in the business of providing a service that is looking to deliver on the promise of the product, which means first deploying AI to help us determine if that promise can be delivered more quickly through coverage assessment." Aspen uses rules-based procedures and models to examine wider data sets, which enables coverage recommendations to be made, but Elliston stresses the importance of all of these procedures being done with a human in the loop who works collaboratively with their AI tools.

Are adjusters adopting AI?

The average insurance adjuster has over 30 years of experience handling a wide variety of claims. Their insights into both policyholders and the claims involved have been honed through years of experience and they see (and know) things most people would miss. The integration of AI into their processes has to complement their daily activities. "We're seeing the greatest adoption in areas where there is clear alignment to a claim handler's day-to-day needs," says Urueta. "Training, transparency and trust are critical. People need to know how to effectively use the tools and be confident with the output. Our claim handling community is directly involved in the design and testing of any AI solutions that we develop – the collaboration not only improves the output but creates a sense of ownership."

Rema Matevosyan, co-founder and CEO of Near Space Labs sees significant adoption of AI in the property and casualty space, particularly within larger carriers. "I'd argue that unlike in other technology waves, the source of the greatest innovation and adoption in the most advanced AI has been the large enterprise companies that are typically characterized as slow to move and resistant to change. They've been uniquely set up to capitalize on this wave, having already invested in the more machine-learning-heavy AI tooling of what I'd call the first iteration of AI, before the onset of GPT-like services."

Miller stresses that "the more a professional understands AI, the more we see an increase in them recognizing that AI doesn't replace their expertise; it amplifies it." He explains that "some of the areas of adoption include photo-based estimating, particularly auto claims, where tools provide fast, consistent assessments without needing to send someone onsite."

AI also provides assistance when it comes to fraud detection, despite enabling bad actors to perpetrate fraud more easily. Miller says that AI can detect anomalies in photos and documentation, making it a valuable tool for adjusters, particularly when dealing with a high claims volume. 

"The use of AI requires a change to culture, roles and skills, legacy practices and processes," adds Elliston. "This is not just about technological change; we are preparing and supporting our people on the development of that technology."

And insurers aren't the only ones using AI for claims. Some enterprising policyholders have used it in the wake of the California wildfires to recreate inventory lists to document their losses for their carriers or to dispute their payouts. While the AI available for this use shows promise, several inquiries into restoration and contents companies found that the technology is still in development. 

Part 2 of this series examines how AI programs are using data and the risks associated with adoption.