Tokio Marine to employ AI analysis for auto damage

Tokio Marine to employ AI analysis for auto damage | Insurance Business

Tokio Marine to employ AI analysis for auto damage

Tokio Marine will harness artificial intelligence (AI) to assess and process auto damage claims across its insurance operations in Japan, in a move that is believed to accelerate claims processing for the company.

Created by London-headquartered startup Tractable, the AI solution will analyse images of car damage, in near-real time, according to a statement from the companies. The data generated will enable Tokio Marine to understand the full range of repair decisions available to it, including recommended repair, paint, and blend operations, as well as the labour hours required.

This, the companies said, can increase the speed of remotely reviewing claims, reducing waiting times from days to minutes by removing inefficiencies from the process. It can also help insurers and repairers to agree on repairs more quickly, getting customers back on the road faster.

“In Japan, after an accident it can take 2-3 weeks to determine the amount that should be paid,” said Tokio Marine group deputy general manager Hidenori Kobayashi. “By using Tractable’s world-leading AI to assess car damage, we expect to shorten that time considerably, cementing our position as a company proud to pioneer new technology that transforms the claims journey for our customers.”

Tokio Marine claims to be the first major Japanese insurer to deploy an AI auto damage assessment solution. It has operations in numerous markets across Asia, Europe, and the Americas.

“We see this as testament to the quality of our technology, the maturity of our solution and trust in our people,” said Alex Dalyac, co-founder and CEO of Tractable. “Computer vision is accelerating accident recovery; the technology is here, it’s on the ground and it’s making a difference.”

According to Tractable, its AI program uses deep learning for computer vision, along with machine learning techniques. The AI is trained on many millions of photos of car damage as well as human appraiser decisions, and the algorithms learn from experience by analysing a large variety of different examples.