Part 3 of Accuracy, Efficiency, and Automation: How AI is Improving Workflows in Property Insurance Claims
Regardless of industry, chances are, you’ve heard of artificial intelligence — otherwise known as AI — by now, and the concept tends to spark the imagination, conjuring images of robots and neon holograms of brains and data points.
In reality, the concept of AI is about teaching a computer to process information like a person, in order to automate certain tasks. In the insurance industry, the major application of AI is within the realm of computer vision which, as it sounds, is teaching a computer to see like a person. The pinnacle of this achievement would be to teach a computer to see and reason like a person, mimicking the abilities of an adjuster or an inspector with greater accuracy and reliability — and at a fraction of the cost and time.
AI, when intelligently developed and effectively deployed, can take on the manual task of looking through the imagery and deriving actionable knowledge. They can perform the pattern recognition tasks of determining, for instance, what is a hip or gable roof, what’s staining versus rusting, what type of shingle is there and how much tree overhang is present.
With companies like Arturo providing the technical development expertise, you don’t need a computer science degree to leverage the power of AI for business decisions. And for claims, this can be applied in a number of ways. The application of AI can cut costs throughout the claims process, reduce instances of fraud, shorten cycle times and reduce operational expenses on loss adjusters. With AI applied to claims alone, loss ratios can go down.
At Arturo, we help you “revisit” a property multiple times a year, and this wealth of historical imagery dating back years can help to more accurately determine when a certain characteristic – a pool, a trampoline – appeared. Having the ability to compare pre-event conditions to post-event conditions during the claims process can help triangulate when the incident occurred and mitigate instances of fraud.
Notably, in one incident in Australia, an adjuster received a claim for a $13,000 pizza oven which, with Arturo’s platform and the breadth of historical intelligence available, they were quickly able to determine had never existed at all, without ever having to set foot on the property.
In addition, before a catastrophic event occurs, granular information about the properties in your portfolio and their resting risk can help to uplift catastrophe modeling. When inputting data into a vulnerability model, to better understand the inherent risk and exposure of a home, using AI-derived characteristics can provide much needed recentness and accuracy. In this way, all estimates of what could happen are based upon a strong foundation of your portfolio’s current health, providing you the best opportunity to be adequately capitalized and ensuring the optimal reinsurance for your book.
In a catastrophic event, you can also overlay event footprints on your book of business to really understand what policies might be affected. By pairing reliable meteorological data with the most recent AI-derived attributes for your portfolio, you can get an “early look” at what policies have the greatest likelihood of being impacted.
For instance, when a hurricane rips through a coastal town, you can see exactly which homes are in the poorest condition and present the greatest policyholder risk and vulnerability within the storm’s path. In this way, you can gauge where you may need resources, both human and capital, so that you can be ready to respond in the areas where it is most needed.
After an event, AI can help facilitate an immediate and effective response. Through imagery providers, “gray sky” flights can be commissioned to capture post-event imagery relating to your portfolio. This imagery is available to Arturo, where our AI models can process the damage severity and loss information and return near time results.
With this approach, you actually know within days – not weeks – of an event who has been affected the most, whether they’ve filed a claim and what their policy entails. Having this information at your fingertips, you can triage claims far more effectively during a national or regional event.
Equipped with the power of AI, now you can understand the severity of an event’s impact and its corresponding claims at a glance. For properties where imagery makes it apparent that damage has visibly occurred, an on-site adjuster may not be needed at all; this gives you the ability to hasten the process and provide peace of mind to an anxious policyholder that their claim is being handled without waiting on a time-sensitive inspection. This enables you to not only optimize time, expertise and costs but also to increase customer loyalty through a better, less stressful claims experience.
Those efficiencies pay off. With the average on-site loss adjuster costing between $400 and $600 per claim, the return on investment on the usage of AI is significant. And because of improved accuracy and reduced rework, Arturo has seen significant improvements on reduced cycle times – an average of 30 minutes per claim.
IAG, the largest general insurance company in Australia and New Zealand, partnered with Arturo to tackle the holy grail of insurance claims: making the claims process as easy as possible. Easier for adjusters, who wouldn’t have to rework a claim or revisit the assessment, and easier for the customer, who could count on a faster outcome, so their home could be restored as soon as possible.
With Arturo’s solution, IAG claims adjusters were able to shave off 30 minutes per claim on average. Across the 152,672 property claims assessed in 2021, that equates to over 76,000 hours or nearly nine years of time saved.
In addition, loss adjusters have utilized this same intelligence to help estimate repair costs. With Arturo’s web application, adjusters can remotely measure roof panels, fences, pools, or anything else that may require an estimate of area and volume to determine potential cost. Across all touchpoints, AI enables post-event activity to go more smoothly.
Now let’s go into what’s next for the future of claims.