How AI Innovation Is Transforming Property Insurance Underwriting

Artificial intelligence (AI) is dominating the technology conversation in every industry—including P&C insurance. In this post, we look at how advances in machine learning, computer vision, geospatial analytics, and other forms of AI are transforming property insurance underwriting in key residential and small commercial lines.  

Key Takeaways

• Thanks to the bar set by digital-first brands like Amazon and Netflix, consumers and businesses continue to gravitate toward P&C insurers that can deliver personalized coverage based on their unique needs

• Geospatial analytics, computer vision, and the rise of multi-modal, AI-based property condition intelligence and automation are revolutionizing underwriting, from quote through onboarding and beyond

• These AI technologies can reduce inspection expenses while maintaining an accurate view of risk—ensuring policyholders get the right level of coverage and reducing persistent levels of underinsurance

• Property insurers will increasingly need to leverage these technologies to enable a shift from reactive to preventative underwriting models that help reduce risks for both policyholder and insurer  

Artificial Intelligence (AI) in Insurance Underwriting

Artificial intelligence (AI) is technology that enables computers to do tasks and solve problems that would otherwise require human intelligence—faster and far more efficiently. In property insurance underwriting, AI is increasingly leveraged to accelerate processes that make use of an ever-expanding universe of data to accurately assess risk and offer personalized, proactive coverage.  

Innovations in computer vision, machine learning, geospatial analytics, and other forms of AI have gained profound importance to property insurance underwriting in both personal and small commercial lines. And it’s easy to see why.

Thanks to the bar set by digital-first brands like Amazon and Netflix, consumers continue to gravitate toward companies that can quickly anticipate and make recommendations based on their unique needs. Perhaps most of all, they prize tech-forward brands that can quickly help them prevent or mitigate issues before they become costly problems. 

Insurance is no different—especially when insurtech challengers and tech-forward incumbents are always just a mouse click or finger swipe away, and even lines like small commercial are moving to real-time quote-and-bind. What’s more, persistent inflation, rising risks, and soaring repair and replacement costs mean P&C insurers are under mounting pressure from every angle. 

As a result, AI is already a game changer for underwriters seeking to write more business, expand into new lines, improve the customer experience, and understand, and price risk quickly and accurately to provide personalized, proactive coverage. Let’s take a closer look at how these advances are actively reshaping insurance underwriting.

The Rise of AI-Powered Property Condition Intelligence

According to McKinsey, insurers employing modern data and analytics technologies can see loss ratios improve up to 5%, premiums rise 10% to 15%, and retention of their most profitable segments climb as much as 10% relative to other companies. 

While Generative AI and other new forms of artificial intelligence seem to emerge daily, putting distance between top-performers and contenders means harnessing AI technologies with a proven track record in insurance underwriting. Case in point: AI-powered property condition intelligence like the kind developed by our team at CAPE. 

Over the last few years, high-quality data on property condition, replacement costs, and other pertinent property attributes have all become widely available. This includes up-to-date intel on the condition of residential and commercial properties captured via aerial and satellite imagery. 

At the same time, the growing expense and reduced availability of inspectors has only grown more onerous. Meanwhile, underwriting in sectors like habitational continues to grow more complex relative to premiums, propelling a need to tap intel that can create new efficiencies.  But while property data grows more valuable and accessible, the effort involved in sourcing and managing multiple data providers and synthesizing a mountain of often contradictory data inputs has grown increasingly unrealistic.  

The consequences of operating without this kind of data, or drawing the wrong conclusions from it, can lead to excessive exposure when quotes are too low—and premium loss when they’re unnecessarily high. In addition to lost business, it can only exacerbate already dangerous levels of underinsurance.

As the industry leader in this space, we recognized a new approach was needed. One that entailed ingesting data from a constellation of public and novel sources, including high-resolution aerial imagery—and then applying geospatial analytics, computer vision, machine learning, and other forms of advanced AI to transform this data into actionable, up-to-date insight, on-demand via API. 

This includes current roof condition and construction, wildfire risk, the presence of a backyard pool, yard debris, or hundreds of other factors impacting risk. The proven loss-predictive power of this intel helps carriers and MGAs improve on traditional data sources for better modeling and stronger business results across a number of use cases, including the following: 

Personal Lines Underwriting: Real-Time Risk Assessment, Instant Quoting

Today, CAPE’s Home Insurance Property Intelligence solution enables The Hartford, Kin and a majority of top US carriers to quickly and accurately price coverage based on an accurate understanding of the risk associated on any of more than 110 million properties nationwide in mere moments. 

This includes a Roof Condition Rating (RCR) that uses computer vision at scale to identify which homes truly require onsite inspection and reduces the likelihood that the premium will need to be adjusted after the policy is already bound. The same risk signals can also identify additional segmentation for new business and renewal pricing models through price-predictive modeling. A roof’s condition, like that supplied by RCR, has been found to be more predictive of loss than the industry’s long-standing proxy, roof age.

Our new Purefill solution combines all of these property insights with public records and data from home sales to make sense of a wider body of unstructured data and pre-fill applications to accelerate the underwriting process from quote. This kind of speed is essential for customers shopping for insurance online and for whom the first to respond with a good rate usually wins. 

Small, Habitational, and Mid-Market Commercial: The Power to COPE Deeper

Commercial property underwriting has long had to contend with inconsistent policy submissions and the complexity of the properties themselves—particularly in habitational and multi-structure commercial where it’s challenging to match the buildings to appropriate policies. 

Underwriters can get tied up for hours looking up imagery from Google Maps, manually geocoding building schedules, or hunting through county tax assessor websites for parcel sketches to get a read on risk. Then there’s the task of translating and normalizing the wide range of formats and data received from agents and brokers to get a read on construction, occupancy, protection, and exposure (COPE). 

But it doesn’t have to be this way. Using solutions like our own commercial property intelligence, for example, underwriters gain visibility into loss-predictive property characteristics for accurate rating and underwriting—automating the workflows they can, and efficiently underwriting the rest with human supervision. 

Available through API, Batch, or using a web application, CAPE’s commercial property intelligence delivers the best COPE data in the industry directly into the appropriate workflows, with support for human review of exceptions and more complex property types. 

And with our new Property Mapper solution, underwriters simply enter the address for an apartment complex or other multi-structure property and automatically get an instant read on which building is associated with what policy—all automated directly into their workflow. 

Reinventing Business Models

AI-based property intelligence also enables insurers to move from reactive to preventative underwriting models that save both policyholder and insurer potentially large amounts of money. 

Under California’s new Safer from Wildfires framework, for instance, insurers must provide discounts to homeowners and businesses that take steps to mitigate wildfire risk by clearing brush and other vegetation away from their property, or address other perils that could ultimately lead to costly claims. 

Using CAPE’s property condition intelligence, carriers can identify true risk to wildfire and other locational perils. Through the solution’s up-to-date change detection capabilities, they can also confirm when mitigation measures have been put in place—adding additional insurer value to policyholders. 

This Is All Just the Beginning of AI in Underwriting

New sources of data, and the ability to apply it through AI, will increasingly become a significant differentiating source of insight and value for property underwriters. In all its forms, AI is rapidly giving insurers the ability to launch new products and services that can make the world more insurable. Put another way: In the $1.56 trillion P&C insurance market, underwriters harnessing the power of AI are very likely to win. 

To learn how AI in property underwriting can help insurers select better risks, reduce expenses, and improve the customer experience, visit CAPEAnalytics.com.