< Back to articles

3 ways AI is causing a revolution in the real estate insurance industry

June 1, 2022

Insurance is all about data. Since its inception, predicting the future and estimating risks have been at the core of the industry. Traditionally, pricing and risk premiums have been calculated based on historical claims and underwriting questionnaires. But with the emergence of new risks like climate change, additional regulations and more demanding customers the industry needs to innovate. New digital technologies such as the Internet of Things (IoT), Cloud Computing, and Artificial Intelligence (AI) can help.

The rise of Artificial Intelligence in insurance

As the Digital Transformation has shown in other industries as well, those that embrace new technologies thrive while those that ignore them falter. New market-entrants like Lemonade underline the need for incumbent insurance carriers to become more innovative and adopt a  leaner operating model. AI allows insurers to do both.

AI makes it possible for insurers to interpret and analyse the seemingly unlimited gigabytes generated by its customers. Data is the new oil and AI allows insurance companies to refine it into usable insights. A clear example of new ways of using data is behavioural analytics which is increasingly used in health care or car insurance. Connected devices like cars or smartwatches allow insurers to estimate and anticipate risk associated with customer behaviour. This type of data enables insurers to vastly optimise their sales, distribution, pricing and claims management.

But still, despite AI's disruptive impact on the insurance industry, it needs data to do so, which in many cases is still lacking. Property insurance is an example of an insurance vertical that has struggled to fuel its analytics with accurate data. Typically it has relied on data that comes from the homeowner or agent, public records, or visual inspections. In many cases, these data sources are inaccurate, outdated, or, in the latter case, relatively expensive. In a data-driven industry like insurance, blind spots come at a cost. Generally, these costs can be divided into three categories: 1) inefficient processes 2) missed revenue or 3) inaccurate risk pricing.

1. Using computer vision to automatically detect high risks

Property & Casualty (P&C) insurance, which makes up about one-third of all insurance premiums, is heavily reliant on manual labor and visual assessment. To insure something physical, like a property, one needs to understand its condition at the time of underwriting, policy renewal, or claim management. All of which depend on time-consuming, manual workflows.

In recent years, artificial intelligence in the form of computer vision has opened up new possibilities to digitise this domain of the industry as well. Computer vision makes it possible for computers to interpret and understand the visual world, allowing insurers to expand their analytics capabilities towards new domains. An example is the automatic classification of the condition of a property or the amount of damage, based on photos or aerial images. Based on this information insurers can see defects like cracked masonry, patched roofs or leakage stains.

2. Detecting underinsurance of properties through images

When it comes to property insurance, customers tend to hold onto their policy until they physically move somewhere else. In the meantime, many mutations can occur to their property that are not included in the insurance coverage, resulting in underinsurance. A recent US study showed that two-thirds of properties are underinsured by an average of 20%, with some homes being underinsured by up to 60%. Accurate and up-to-date data can support insurers in detecting underinsurance and offering the appropriate coverage to their policyholders.

Computer vision can help insurers to do so. For example through the automatic screening of property aerial images, which can provide details about the property such as its size or the risk of calamity. Such analyses even allow for remote valuation of the property to determine the reinstatement value. The outcome of these automatic screenings can be used to detect any deviation between the policy information and the latest state of the property.

Aerial and street-level images also allow for creating visual time-lapses that show changes in a property, or its estate, over time. Information like this can serve as a visual baseline in case of a claim for example. It also allows for proactive action from the insurer toward the policyholder to nudge different behaviour and prevent risks.

3. Improving the calculation of risk premiums

Digital innovation provides insurers with new ways to underwrite traditional risks, often by using individual rather than group data. Personalised data allows for custom pricing of insurance. This helps carriers to improve their loss ratio and incentivises the insured to reduce their risk of calamity. An example that has been given earlier is the behavioural data of policyholders that can be collected through connected devices. Also for property insurance, it is possible to obtain such personalised data to offer a tailored product.

Computer vision allows insurers to automatically verify the age, condition, and characteristics of a property, as well as its potential for hail and wind damage. This supports insurers to accelerate their underwriting process and assess future risks with higher accuracy. With high-resolution images and aerial imagery, it is possible to capture potential hazards, such as nearby or overhanging trees, materials that are especially prone to damage, and expensive attachments, such as solar panels.

Recent years have also shown an increased risk of climate-induced damage caused by wildfires or floodings. As these risks grow, computer vision helps insurers to measure property elements like elevation and acreage, as well as monitor space between structures and vegetation or other potentially combustible materials for wildfire mitigation. The combination of these characteristics allows for the creation of property-specific hazard scores that summarise the overall risk of natural disasters.

Want to know more?

Here at Spotr, we use geospatial imagery to automatically inspect the condition of the property and its associated risks. Spotr is an AI-powered remote property data platform, helping you to gain insights into your insured portfolio. Spotr uses computer vision to analyze a huge, up-to-date image database in which data about your real estate portfolio from different sources are collected. With this enriched database, you can easily analyze your property insurance portfolio.

Want to know more? Contact us here: