Car crashes are an unpleasant experience, and not only for the people involved – but for insurance companies too. Determining the extent of the damage suffered by a vehicle and the cost of repairing it requires experts to inspect the car or view pictures of it. But such specialist are not always at hand, and it can take days before a claim is settled.

In late 2014, Alex Dalyac and Razvan Ranca, two AI experts who had paired up through London-based company-builder Entrepreneur First, joined forces with e-commerce entrepreneur Adrien Cohen to start a business fast-tracking car accident settlements after realising the sector was in need of a shake-up. Furthermore, Cohen says, “it's first and foremost visual: people assess accidents by looking at photos of damaged cars.” AI algorithms using deep learning can outperform humans at visual assessments if provided with enough examples. The trio asked insurance companies for access to their massive archives of car damage pictures, and hired automotive engineers to help classify the images to train their AI, before patenting a specific tool to speed up labelling.

Fast forward six years and their startup, Tractable, has grown into a 100-employee business operating in eight countries, and partnering with insurance companies including Covea, Ageas and Tokio Marine. When customers of Tractable’s partner insurers have a car accident, all they need to do is take ten pictures of their vehicles. “The photos are collected, they're sent to the cloud where our AI will process them – and instantly we will send the full assessment to the insurance company,” Cohen says. “The claim handler can then send a car through the right workflow or offer a cash settlement.”

Next is taking the service to more countries – Cohen says he is aiming for 30 “in the next few years” – and into other sectors, such as assessing damages to buildings following earthquakes, flooding and other natural disasters. “That obviously is a natural expansion [of what Tractable does], and the insurance companies have the photos,” explains Cohen. “It's not a matter of if – it's a matter of when.”

Gian Volpicelli is WIRED's politics editor. He tweets from @Gmvolpi

🏘️ Failing care homes are the real coronavirus scandal

🔒 The UK's new lockdown rules, explained

❓ The UK's job retention furlough scheme, explained

💲 Can Universal Basic Income help fight coronavirus?

👉 Follow WIRED on Twitter, Instagram, Facebook and LinkedIn

This article was originally published by WIRED UK