This piece is by Omer Perry, Product Manager at Ravin AI and looks at the benefits of using AI to examine images during the claims process.
Insurance is one of the last holdouts of the pre-AI era, with claims still reviewed and settled based largely on the results of investigations by adjusters who come to the scene in the aftermath of a fire, storm, or vehicular accident. It’s a time consuming process that delays and complicates settlements – and the lack of real-time data during an event leaves insurers vulnerable to fraudulent claims; a problem that gets worse every year.
The current system does not properly serve customers or insurers – which means that it is ripe time for a change. In the past, technological limitations forced both customers and insurers to rely on an inefficient claims and investigation model, however, that’s no longer the case; by utilizing advanced computer vision technology, insurers can get accurate, up-to-the-minute details on events.
For example, mobile phone cameras could be used to document damage immediately after an incident occurs, enabling the insurer to get the data it needs to begin processing and evaluating a claim. With the images as fresh as possible, insurers will get an accurate picture of the damage, reducing the likelihood that the scene will be tampered with. And with the images transmitted back to the company’s artificial intelligence-equipped servers for analysis, companies can quickly determine the true extent of damage, how much it would cost to repair, and determine what customers are entitled to, without forcing them to wait endless weeks for answers—or their money.
Such an approach can work in various types of insurance categories, and will help retain customers and reduce costs, which are key to success now as inflation continues, and policy prices rise.
A path to better evaluation of structures before and after claims
A fire or major storm can completely gut a home, commercial building, or factory – but how much is that damage worth? Adjusters sent out to the scene of an incident take photos and record what data they can, but they are unlikely to get a full picture of the damage; for example, in specific cases, some areas are just too dangerous for them to tread. Computer vision can play a key role here: A drone equipped with a high-resolution camera can navigate over, under, or inside the damaged structure to collect images. These images – along with other information, like data collected by connected in-home devices, such as smart appliances or thermostats – can be transmitted immediately to the company, where it will be analyzed by advanced AI systems.
In addition, computer vision technology, deployed via mobile phone cameras and apps, can help insurers determine the true state of a structure when issuing new coverage policies.
For example, a structure can be scanned with a phone camera, and its exact condition documented and analyzed by AI in an objective way. This allows insurers to better rate the condition of the structure when underwriting a policy, taking on itself the appropriate level of risk – and paying out a settlement that is fair to the customer and to the company if and when the time comes.
A more clear picture of vehicle conditions and damage
Accidents and incidents involving vehicles are usually difficult to evaluate – and are the source of most complaints and settlement disputes by customers. Insurers have to rely on reports by customers and, sometimes, police, in order to determine if a claim is possible – after which they dispatch an adjuster, who usually only gets to see the damage when the vehicle is already slated for repairs. And while adjusters are good at weeding out a lot of potential fraud, data shows that fraudulent auto insurance claims alone account for more than one third of fraud losses by insurers. Checking claims for fraud, a common practice at most insurance companies even though most states today have no-fault insurance, often lengthens the claims process, meaning that customers need to wait longer before getting their payouts.
Here, too, advanced computer vision and AI can help both insurers and customers get an accurate picture of damage – and reduce the likelihood of fraud. Customers can file an accident report on the spot using the insurance company’s app. A customer or insurance company representative can also immediately use a cell phone camera to capture images or videos of the car, and upload it to the insurer’s system.
Using its large database, the insurance company will be able to analyze the scene of the accident, as well as the condition of the vehicle, to determine responsibility and risk factors – whether the driver was speeding, distracted, or otherwise partially or wholly responsible for the incident. If a customer was indeed the victim of another driver, or of circumstances, the company can establish a payout immediately. And at the very least, the company will be able to avoid fraudulent claims, as the data and images gathered within minutes of the accident will provide an accurate picture of what really happened in real-time.
Like in the cases of structural damage from storms or fires, AI can also be used when issuing a new insurance policy; as AI can evaluate images of a vehicle taken with cell phone cameras or other devices to help determine that car’s baseline value and condition. Not only is this often more efficient, both in the onboarding and claims process, but it also creates a more objective record and avoids human error and oversight.
While such uses of computer vision and AI are not yet mainstream, the technology exists, as do the systems to implement it. Now insurers must work toward adopting it.