Artificial intelligence, machine learning, and automation continue to transform countless industries, and the world of insurance is no exception. As the benefits of progressive technologies become increasingly clear to companies and customers alike, we can expect the entire risk management field to start capitalizing on these opportunities.
While the phrase artificial intelligence has other meanings in popular culture, for our purposes it refers to machine learning processes that analyze patterns in massive volumes of aggregated data. This lets insurance companies use customer data to generate predictive pricing models far more accurate than traditional algorithms. Businesses within this field also benefit from AI’s automated potential, enabling agents to focus on client interactions instead of monotonous number-crunching tasks.
Customers can benefit from AI-powered insurance as well. Using precise customer data, machine learning systems can generate highly personalized plans and premiums that reflect the client’s history – lowering overall costs. Automotive insurance plans will analyze driving history when offering rates to drivers, and health insurance appraisals will reflect everything from family medical history to diet and exercise. All risk management firms use algorithms that calculate risk, but machine learning examines more comprehensive data sets. Digital tools also make it far easier to collect personal information and spot fraud risks.
Though machine still relatively new, companies are already innovating with it. Lapetus Life Event Solutions drew attention by requesting selfies of life insurance applicants. These images were processed by AI software to generate risk comparisons against groups such as smokers without requiring a medical exam. Meanwhile, the SVP of Insurance for Sutherland has proposed wearable sensors that monitor the health of disability claimants. London-based Neos Ventures offers lower home insurance premiums to customers who allow sensors and monitoring software installations on gas and water lines. With this technology, Neos gains accurate information on product performance and can contact repair or emergency services in the event of a problem.
Most of these scenarios require some kind of privacy trade-off, and so far customers seem to be split on this. One study found that almost 50% of consumers were willing to provide biometric data through wearables for reduced health premiums. But in the car insurance market, 21% of customers refused to participate in any usage-based insurance monitoring. Of that group, 81% didn’t want to be monitored, doubted premiums would decrease, or didn’t believe they’d save money. That said, constant monitoring means insurance companies must implement robust digital security measures, especially when customers file claims through online platforms.
While AI insurance has many benefits, the true appeal may lie in claims settlement. In 2017, the insurtech startup Lemonade famously claimed that its AI resolved an insurance claim in less than three seconds. Even if the company exaggerated this statement, AI can heavily optimize claims processing, especially when submitted through digital platforms. This is a crucial point for customers, who often cite claim settlement speed as the most important factor in rating an insurance company.
Fast settlement claims are great for companies, but reduced fraud risk is even better. AI optimization allows insurance providers to resolve inconsistencies in claims paperwork far quicker than human agents can. That puts AI in an ideal position to spot false claims, typically created by identity thieves. It’s currently estimated that AI could ultimately save insurance providers $80 billion in fraudulent claims, a bottom line issue for most companies.
Some insurance companies have already started integrating AI solutions into their platforms, though the process is not without obstacles. The 2017 Excellence in Risk Assessment report noted a general lack of awareness of emerging technologies from most legacy insurance players. Those who have implemented machine learning often waver on how to apply the data to pricing models.
This is understandable – insurance underwriting often relies on historical data that reflects the risks of digital landscapes. Cyber fraud, account hacking, and virtual privacy breaches are relatively new from the perspective of risk management. AI could help with predictive modeling to provide a relatively accurate rating of new risks, but insurance companies still need to make that leap before realizing the benefits.
Perhaps a bigger concern will be the challenge of inspiring trust in the insurance industry itself, AI or otherwise. The fact that insurance companies already have a negative reputation when it comes to unreasonably declined claims is well known — in Australia, more customers trust sex workers than insurance agents. Customers may carry inherent distrust even when it comes to new tech-based insurance solutions. One study found that roughly 60% of respondents weren’t interested in purchasing insurance from a chatbot, which means that human agents would still be required in many scenarios.
Despite these challenges, AI is likely to be a major driver of innovation and technological change across the insurance field in the years ahead. The good news, is we’ll start seeing more industry winners as the technology is refined – including the customers themselves.
