Residents of Southeastern states like Florida will be looking ahead to the heart of hurricane season with trepidation. With climate change bringing storms that are wetter, slower and more prone to intensifying rapidly, insurers are struggling to assess risk and underwrite profitably, and customers are struggling to afford skyrocketing premiums. In 2022 across the U.S. there were 18 climate disaster events, including floods, severe storms and wildfires, the third most destructive year ever. Florida, in particular, suffers from hurricanes, with six of the 10 costliest storms in American history hitting the Sunshine State.
Problems have been exacerbated by Florida’s excessive amounts of litigation and claims fraud. In this fraught climate, insurers are managing climate-change risk and reducing fraud by using aerial imagery and artificial intelligence to extract insights about a property’s risk profile, so they are better able to detect fraud and price premiums.
Carriers now have the ability to leverage data from historical imagery while assessing insurance claims, enabling them to detect and prevent fraudulent activities. By analyzing such data, carriers can determine when the damage happened, working out whether it predated the insurance policy or occurred after a natural catastrophe. Additionally, historical imagery data can reveal the extent of the damage and whether a specific event exacerbated the severity. This valuable information helps carriers make accurate assessments and reduces the potential for fraudulent claims.
In the past, insurers used historical trends to assess the risk of a claim, but with erratic weather patterns they are now giving a heavier weighting to information about a property’s location, roof age and condition, flood risk, first-floor elevation, and proximity to vegetation if located in a wildfire risk zone. Across time, they can then start to build up intelligence about trends in different areas to see, for example, whether flood levels have increased.
For years, gaining this information required insurers to send out inspectors to residential or commercial properties. But today, accessing valuable insights is cheap and immediate. Underwriters then have more information to price premiums competitively. They can also proactively reach out to customers, telling them how they can decrease their risk, for example by cutting down overhanging trees close to their property.
One property intelligence phenomena we have seen in recent months is the use of drones to collect detailed data in areas that are difficult for humans to reach. However, drones remain an expensive option as they still require a human operator, meaning this type of data collection is not cost-effective. The real advance that is helping the industry improve their property insights at a massive scale is improvement in aerial imagery from aircraft. Resolution is now so accurate that details at a scale of 7 centimeters can be collected across huge areas. As a result, we are seeing many more insurers use aerial imagery and AI to collect property intelligence. We have grown our own client base by 120 in 2022.
After a disaster has struck, claims processing can be streamlined, with fraudulent claims detected quickly. Rather than relying on individual teams to physically inspect damage to a property, insurers can use aerial imagery or even footage from street cameras to assess post-disaster damages. By using pre and post-disaster imagery, insurers, with the help of AI, can remotely examine the extent of damage to a property and identify exaggerated claims. They can also determine whether damage occurred due to the weather event or if it already existed on the property. For example, insurers can access evidence about roof wear, rust, or whether a tarp was already on the roof indicating pre-existing damage.
In particular, it is advances in AI that have caused a great leap forward, particularly in the field of predictive analytics for underwriting and claims. Because computer vision is able to analyze larger sets of imagery than ever before it is now possible to create and extract new types of attributes that were previously impossible to track. This has meant that suddenly it is possible to view aerial imagery at a large scale. For example, an insurer can now take historical imagery from 20 years back and identify a change like a roof replacement automatically. Doing this manually would be too expensive, but AI makes this level of tracking feasible for the first time.
With every year, imagery analysis becomes quicker and cheaper, meaning insurers can now create data sets that allow them to respond in the most effective way to a natural disaster and allocate resources in the most efficient way.
The challenges facing the insurance industry in the face of climate change are significant. Floridians are facing a projected 40% hike in property insurance premiums this year. At the same time, technological innovations offer some hope. By plugging data gaps using artificial intelligence and aerial imagery, insurers can viably continue protecting Americans from having to foot a life-changing bill post-disaster.
GeoX is a data analyitics company that uses machine vision and deep learning technology to extract information from aerial imagery for insurance companies.
Was this article valuable?
Here are more articles you may enjoy.