Drones, algorithms and smartphones… powerful new risk assessment tools

Aerial resources

After the passage of Hurricane Harvey, insurers based in Texas deployed fleets of drones to inspect sites. Farmers Insurance estimates that by using drones, an adjuster can process three damaged houses in one hour.

To minimise the stress on its Texan policyholders, the automobile insurance specialist Esurance resorted to predictive analysis. Shortly before the hurricane hit the region, the company booked several aircraft, which are more versatile than drones for image acquisition. As soon as air traffic resumed, this fleet of aircraft was mobilised for photographic reconnaissance missions in flooded areas.

Once these data had been collected, algorithms estimated the degree of damage to the policyholders’ vehicles according to several criteria, such as the depth of water in which they had been damaged, as seen from the sky.

Estimates by MMS picture messaging

After an initial experiment in 2016, which included the analysis of several thousand photos of damaged cars, the British subsidiary of Belgian group Aegas now uses Artificial Intelligence to assess the repairs required, based on simple photos. This technique also improves fraud prevention. As soon as the machine spots an anomaly (on photographs or in a notification), it reports it and a claims manager then takes over.

This tool, named "AI Approval”, has been developed by the Fintech Tractable. Claims can be processed in just a few minutes. This frees up time that adjusters can devote to more complex cases.

Identical initiatives are proposed within the Crédit Agricole Assurances Group.Pacifica has been offering a claim management application since 2011. The main functionality – “Submit a claim” – enables policyholders to enter the circumstances of the accident directly and attach up to five photographs in support of their notification. This notification procedure is fully integrated into Pacifica’s processes, guaranteeing the optimal handling of the insurance claim.

Our Italian subsidiary Crédit Agricole Assicurazioni has followed in Pacifica’s footsteps by launching its own application last year. Policyholders can submit their notifications of car insurance, personal cover and home insurance claims.

Watching out for fraudsters

Closer to home, Shift Technology – a French start-up founded in 2013 – is enabling insurers to detect fraud more quickly.

Using machine learning, a technique based on Artificial Intelligence (AI), the start-up has already reviewed over 86 million claims for all types of losses (automobile and home) on behalf of its clients. The detection algorithms developed (based on the expertise of anti-fraud experts) can identify fraud with a 75% success rate.

They are also capable of identifying the most difficult type of offences to detect: organised fraud. However, this tool is limited to alerting managers about fraud, who can then proceed with more detailed investigations.

It is interesting to note that these innovations have emerged in three different countries simultaneously : in France, United-States and England. They share the characteristic of supplementing the adjuster’s role rather than replacing them, and they save significant amounts of time. They are a tremendous success with policyholders. And this is just the start: certain companies are even contemplating using predictive models to compensate policyholders before hurricanes strike!