As inflation causes the cost of repairs, labor and benefits to rise, insurance premiums also rise across the board. Auto insurance premiums have risen 13.7 percent nationally over the past year, according to a Bankrate.com study. Home insurance, meanwhile, rose 12.1% year over year, Policygenius found.
But Jonathan Matus argues that it doesn’t have to be. He is the founder of Fairmatic, a company that applies AI to – at least according to him – reduce risk in the auto insurance industry.
Matus previously founded Zendrive, a platform that provides information to businesses for auto insurance and claims, as well as roadside assistance. While Zendrive focuses on insurance for individuals and families, Fairmatic has a more commercial bent – a customer base made up primarily of businesses.
“Having spent about a decade of my career at Google and Facebook, I quickly saw the negative externalities of the technology I actively helped make widespread,” Matus said in an email interview with TechCrunch. “The path to creating Fairmatic was created by the need to reduce the risk of one of the worst externalities of this powerful technology: the increase in distracted phone use while driving and the consequent loss of life on the road.”
Matus may speak in grand words, but Fairmatic’s business proposition is simple: analyze and price the risk profile of the fleet. The company uses AI models trained to manage data to try to reduce risk and help with various policy and claims management processes.
Customers get access to an app they can use to monitor “driving events” – such as erratic driving – and “identify actionable improvement opportunities”. The app also offers what Fairmatic calls a “fully digital mobile claims experience” that can automatically detect crashes (hopefully better than Apple) and analyze incident data.
Here’s Matus: “With Fairmatic, fleets of small, medium and large businesses are empowered with actionable insights that improve safety and have a direct impact on insurance savings.”
But there are reasons to be skeptical. Fairmatic isn’t the first to bring AI to auto insurance decision-making—Jerry, Just, Root, and Tractable offer similar technologies, albeit consumer-focused ones—and AI has a spotty track record in the insurance industry.
Last year, the Casualty Actuarial Society (CAS), the professional association of actuaries specializing in property and casualty insurance, acknowledged the harmful effects AI can have when used by financial institutions in determining insurance and mortgages. In a series of papers, CAS concluded that biased data – data about which insurers train their algorithms – could perpetuate discrimination that already exists in the insurance industry. (See: Allstate’s Pricing Algorithm That Disproportionally Affected Nonwhite Customers.)
A subsequent report from the California Department of Insurance highlighted particularly problematic recent uses of AI by insurers, including tagging claims by city zip codes and using non-risk-related personal information in the marketing and underwriting of insurance policies. “Conscious and unconscious bias or discrimination … can and often does result from the use of AI as well as other forms of ‘big data,'” the report’s authors wrote.
Washington and Oregon tried to ban the use of credit-based algorithms to determine auto insurance premiums. Separately, Colorado introduced legislation requiring insurers to test their algorithms and scoring models to uncover bias.
Matus is adamant that Fairmatic takes care to reduce the potential for bias in its AI. In fact, he argues, Fairmatic’s reliance on AI has generally resulted in better outcomes for customers who have historically been forced with insurers using outdated data and pricing models.
“Fairmatic’s AI predictive risk model is trained using over 200 billion miles of driving data,” he said. “This allows us to develop a deeper and more holistic understanding of the unique risk profile of each fleet and each driver, and then distill that data into insights and training that improve driving behavior and reduce risk.”
But even if it’s true that Fairmatic’s approach is better than most, the platform’s driver monitoring capabilities are worrying in their own right. They are reminiscent of the algorithms that Amazon has used to monitor the behavior of drivers during the workday. According to Vice, the algorithms improperly penalized drivers when cars cut them off — data that Amazon used to rate driving performance and determine individual bonus payouts.
In response to a question about monitoring and data privacy, Matus said that Fairmatic “only actively monitors data related to the insurance policy,” including information necessary for risk and claims management. “Fairmatic’s technology layer uses anonymized data and does not retain driver information without fleet permission,” he added.
Whatever the case, Fairmatic has had no trouble attracting investors — or customers, for that matter. The startup raised $46 million in a funding round led by Battery Ventures with participation from existing investors and Bridge Bank that closed this week, bringing its total funding to $88 million, double its previous valuation (Matus would not provide a figure). . As for customers, Matus says it has onboarded “hundreds of thousands” of drivers.
This is not very surprising. Commercial auto insurance is a huge market, with Allied Market Research estimating it will be worth $307.10 billion by 2030 – up from $128.44 billion in 2020. Risk of bias aside, money matters a lot – even at a time when global insurance VC funding continues to decline.
Fairmatic’s short-term plan is to hire 30 employees for its research and development center in Israel and more in Bangalore, India, Matus said. The company’s current workforce – spread across offices in the US, Israel and India – totals 85 people.
“Fairmatic has seen tremendous traction since our Series A and this new capital will be used to accelerate our leadership in bringing the full power of AI to commercial auto insurance. Fairmatic is making strategic hires worldwide and we are accelerating the growth of our R&D centers in Israel and in India,” said Matus. “Our vision is to build a technology-driven insurance platform that is not only fully digital, but also comprehensively enhanced and powered by AI, across every product and business function.”