Global climate catastrophes have been growing dramatically in number and severity this century, as 2021 saw natural disasters cause worldwide losses of $280 billion, the fourth-costliest ever. Tractable, a London-founded developer of artificial intelligence systems for visual assessment and disaster recovery, claims it can drastically shorten the payment process through its new AI Property, which uses artificial intelligence to assess the amount of external damage to buildings caused by wind, hail, hurricanes and other large natural disasters.
Global climate catastrophes have been growing dramatically in number and severity this century, as 2021 saw natural disasters cause worldwide losses of $280 billion, the fourth-costliest ever. Not all homeowners have insurance to address their losses, but even those who do often face a lengthy process until coverage payment is received from an insurer.
However, in the latest example of insurtech intersecting with proptech, Tractable, a London-founded developer of artificial intelligence systems for visual assessment and disaster recovery, claims it can drastically shorten the payment process through its new AI Property, which uses artificial intelligence to assess the amount of external damage to buildings caused by wind, hail, hurricanes and other large natural disasters.
AI Property is designed to accelerate the damage assessment timeline from several months to a single day, said Giacomo Mariotti, senior product manager for Tractable It has been in use in Japan since last year, through Japanese insurer MS&AD Insurance Group. The AI helps homeowners in Japan recover faster after natural disasters such as October 2021’s Typhoon Mindulle.
Tractable launched the AI in the U.S. in January and is pursuing partnerships with U.S.-based insurance companies, Mariotti said. “We work with insurance companies [as a ‘white label’] and we don’t appear to the policyholders. It’s all done through the customer experience of the insurance company, which manages the relationship with the [homeowner].”
Owners can access AI Property through a mobile-friendly web-based app on their smartphones to take photos of damage from major climate events, Mariotti said. The AI, which has been trained on a large database of claims and damaged property, makes an immediate assessment of the amount of damage it sees and relays this to the homeowner’s insurer.
“At a high level, this is about accelerating property insurance claims,” he said. “We do this by providing access to an expert appraiser for policyholders by tapping on their smartphones. The photos are used to appraise the damage and help speed up the resolution of a claim by giving the insurance company an understanding of the damage without having to send someone on the ground to get this information.”
However, the science and practicality of disaster mitigation AI is still being developed, Mariotti said. “Something that is quite important is how do you take photos of things that the policyholder can’t reach, like the roof, for example?” he said. “On this one, we are working with aerial imaging data providers to provide such information, as well. It is not captured by drones in our case, but by satellite or fly-over images captured by planes. The reason is that a drone requires a drone pilot to get to the property. So you still have to rely on someone that goes to the property for that specific home. That is really difficult to do when many properties are affected in the same area at the same time.
“Nowadays there are programs run by, for example, the Geospatial Insurance Consortium, that right after a catastrophe hits can get planes to capture imaging within 24 to 48 hours and get the imaging at scale for the affected area. So I think that’s one way we found very effective to be able to process claims at scale.”
For now, Tractable cannot completely verify claims and prevent fraud through its technology, but they are working on the issue, Mariotti said.
“At the moment, we are not doing this type of fraud check,” said Mariotti. “It’s on our [technology development] roadmap. The good thing about using a modern smartphone to take photos is that we record some metadata information, like the location of where a certain photo is taken. So we envisage using this type of information to double-check if the location corresponds to the address of the homeowner.”
At this point of development, Tractable’s AI cannot assess interior damage caused by natural disasters due to the myriad item and pricing complexities involved, Mariotti said. “There are a lot of interior damages that may not necessarily be driven by disaster. They may be a broken pipe, but they still affect people’s livelihood in their day-to-day life,” Mariotti said. “We think that empowering policyholders in these difficult moments can help them resolve the problem faster,” but the AI still needs to be taught a vast amount of information to mediate interior damages. We are working very closely with expert appraisers to teach our algorithms how to do things,” Mariotti noted about the exterior and interior AI-building process.
“We leverage human expertise to build these algorithms and we leverage a lot of historical data about claims that have happened in the past to learn the patterns of property damage,” Mariotti said. “The main reason behind wanting to have an AI driving this process is consistency and scalability. An algorithm-based solution can deliver instantaneous appraisal, even if the volume is much larger.”
Sources:
Commercial Ovserver
Provided by the IKCEST Disaster Risk Reduction Knowledge Service System
Comment list ( 0 )
