Artificial intelligence is on the rise in insurance, and Metromile, a usage-based insurance carrier that charges customers based on how much they drive, is trying to get ahead of the pack.
The company launched AVA, an AI platform, in July, billing it as a “smart claims assistant” that can verify claims information based on the data from Metromile’s telematics device, Pulse, and get customers’ cars into a repair queue faster. It has since added significant capabilities, including routing cars to a specific repair shop, renting a loaner vehicle and making claims payments automatically, without customers having to get involved.
Sean Griswold, chief customer officer for Metromile, says that AVA is more than a dressed-up rules-based system. The company has employed data scientists and engineers to build the system from the ground up, integrating it with all components of its value chain including the traditional policy administration, claims processing and customer service platforms.
“The machine learning and algorithms that support the system is certainly far beyond rules-based, the tech that we’re building in terms of how we interact with the consumer and how we learn from those interactions isn’t rule-based,” he says. “We’re definitely digging in deep on this.”
AVA has been under development from a programming perspective for about three years. Access to a wide range of data sets is crucial for AVA to deliver on its promise to “delight” customers.
“We absolutely use the data from the entire consumer lifecycle in building the learning models and AI,” he explains. “That includes telematics, underwriting information and interactions the consumer has with us pre- and post-claim.”
The commitment to true AI at Metromile is a top priority for the company, Griswold adds, as it aims to compete on customer experience.
“The insurance business is pretty manual in its processing, and customers don’t get the best experience possible,” Griswold says. “It’s not just about moving forms online, it’s about automating the system.”
