AI Takes Center Stage in Insurance Pricing

The use of artificial intelligence (AI) is expected to have a significant impact on the auto insurance sector.

Complex algorithms will help better price risk and set up profitable pay-as-you-drive business models. Insurers are also finding ways to improve accident claim processing from start to finish using AI because it promises to be used to more quickly and accurately spotting inconsistencies in claims or premium assessments.

Stockholm software company Greater Than develops AI software that claims to price risk per car in real-time. Its founder Sten Forseke, said this could “revolutionize how carriers price risk and provide new business models through behavior based and use-based pricing.” He claims the company has built a database of 750 million unique driving situations and has trained the AI with real claims or situations.

The AI then makes a DNA of the trip and that data is sent to the database, which looks for patterns that have led to accidents and, from there, risk exposure can be calculated. “Our AI makes 4.1 billion decisions a day, which means the learning curve for AI is so steep as to be almost unbelievable,” Forseke said. “It is very suitable to help humans with quantitative analytical processes. The potential is enormous.”

Tractable’s AI is trained on many millions of real images of car accidents and repair operations and claims to assess auto damage as a human appraiser would, explained Cornelius Vogel, its head of the DACH region in Europe. Insurers can use this to evaluate the damage to a vehicle, based on photos provided by repairers, appraisers or policyholders. The technology shares repair method recommendations and guides the claim process to ensure each claim is processed and settled as quickly as possible.

Vogel said by applying their AI to claims, it takes just minutes to assess damage and create estimates, instead of days or even weeks and images uploaded by the policyholder can be processed immediately following an accident. “Whenever a recovery of a motor claim starts, it starts with a visual appraisal and this needs to be an expert,” Vogel said. “Our software, trained on hundreds of millions of data points, can estimate repair costs and damage, essentially combining the expertise of thousands of experts.”

Vogel noted, on a general level in the auto insurance industry, the core element in motor claims is the visual assessment, so it is in this respect that the entire industry is looking to see if AI-enabled computer vision is at a stage where it can be deployed. “We focus on the quality level of the claims assessment and the speed we bring to it,” he said. “We keep telling insurers ‘Test it live’ and they can immediately see that it’s very advanced already.”

Vogel pointed out AI helps the auto insurance industry with underwriting, allowing them to combine the fields of claims data, telematics data, resulting in more precise underwriting in the future. That extends to underwriting of used vehicles, which he points to as an area where underwriting policy has been particularly tricky. He claims AI computer vision can more accurately assess the condition of a pre-owned car and combine that data with make, model, year and other metadata to provide a more accurate assessment for the insurer.

The technology could also aid insurance companies in reducing fraud in claims by being able to detect manipulated imagery, Vogel noted, and can also help streamline operations by moving human resources to areas where it can be best used. “AI will process all images automatically and experts can shift their time allocation to more complex areas, which will impact the way firms will use their motor experts,” he said. “Insurers are looking for a full digitalization of the claims process, from the first notification to the repair shop and so AI will have a big consequence, not only for large insurers, but for small and medium-size companies as well.”


Leave a comment

Your email address will not be published. Required fields are marked *