By AI Trends Staff
Technology-driven insurance businesses – insurtechs – are startups helping established insurers study how to gain an advantage by employing AI.
Take the weather. It’s been unsettled recently, forming new patterns, reaching new extremes. To gain insight into changing weather patterns, insurance companies are turning to AI. A recent analysis by Deloitte cited in an account in FinTech stated, “Advanced analytics could further help companies assess historical weather records, insured property data, and assumptions regarding future climate conditions to improve risk selection and pricing.”
AI acting on data from a growing number of Internet of Things (IoT) devices is allowing companies and scientists to better track and understand global weather patterns.
Understory of Madison, Wisconsin, is a startup at the forefront of the merging AI and IoT for insurance companies. The company makes an on-the-ground weather station equipped with sensors – called Dot – for measuring hail storm intensity. Dot is designed as a steel ball on top of a column set into a base, about four feet high in all.
The National Oceanic and Atmospheric Administration documented more than 4,600 hail storms last year causing billions of dollars of damage. Car dealerships with open lots can be especially vulnerable.
Dots are currently installed in six US cities, collecting data shared with several major insurance brands. One of those companies is reported to have saved up to 20 percent in claims by improving claim accuracy after storms.
Founded in 2012, Understory aims to reduce insurance costs for auto dealers with its device that gathers data from many weather conditions, including rain, wind and hail. “Customers finally have access to simple and cost-effective hail coverage,” stated CEO Alex Kubicek stated in a story from the Wisconsin State Journal. “This is a truly disruptive way of doing business in a very traditional industry.”
The Understory sensors automatically begin a claims process when the Dot sensor on site detects a storm above the policy’s threshold. The data from the sensor is used to estimate the amount of damage resulting from the storm. Dealers can potentially receive a payment on their claims within two weeks. Five years of weather data helps the company assess risk and determine insurance premiums. “It’s all based on the data,” Kubicek stated.
Drones in Use for Years
The insurance industry has employed aerial drones for years. Leading insurance companies sought permission from the FAA as early as 2015 to use drones to collect data, aid in catastrophe response, conduct research and development, and support underwriting and claims resolution. A recent report from Deloitte noted the arrival of startup insurtech firms offering specialized services to the insurance industry.
Betterview, an insurtech using drones for property inspections, has conducted more than 6,000 rooftop inspections in the last two years. The company recently signed an agreement with Loss Control 360, which provides software for insurance companies and inspectors. UK-based ingenie aims to help young drivers develop good driving habits, and qualify for insurance savings.
Auto Insurers Employing Telematics, Natural Language Processing
In car insurance, according to a report from insurer Lloyd’s, the use of telematics to collect driving data from vehicles in real time, is causing some disruption. Data from IoT sensors combined with machine learning algorithms can create personalized risk profiles for drivers. Insurers are using the data to give discounts to drivers with safe driving habits, and penalize unwanted behavior such as speeding, hard braking, harsh acceleration and hard cornering. Data can also be used to reconstruct accident scenes to enable insurers to better understand what happened, and process claims more quickly.
In the US, major insurers offer set discounts by state for use of telematics. A recent account in The Zebra lists the discounts, ranging from $0 to $300 per year in savings.
Lloyd’s operates an accelerator program, called Lloyd’s Lab, which has spawned Layr, a cloud-based commercial insurance platform that uses AI to assess the risk of insuring small businesses, according to an account in The Next Web. The system aims to match applicants with the right policies. Layr is based in Atlanta on the campus of the Georgia Institute of Technology.
In health insurance, data collected from wearables such as fitness trackers and heart rate monitors enable insurers to reward healthy habits such as regular exercise, and encourage preventive care with healthy nutrition tips, for example. One insurtech startup is Collective Health, which works to match its members with the right healthcare options by using machine learning to help assess risk. The AI models combine claims data, prior authorizations, eligibility data, engagement data and healthcare use data to create a profile of each member. It can help identify a member who might be confused by prescribed medications and needs to speak to a pharmacist, or identify members who might need home healthcare services.
“AI reduces the task of manually reviewing thousands of medical claims, and instead focuses our staff on performing warm, human outreach, and thinking through complex problems together with our members,” stated Dr. Sanjay Basu, Collective Health’s Director of Research and Analytics.
Exploiting Natural Language Processing Advances
Improvements in natural language processing are also being employed by the insurance industry to process unstructured text, in the form of documents, chat logs and emails generated in day-to-day interactions. Insurers are able to leverage language models to reduce the time it takes to respond to customer queries and find relevant information, from the many documents they must review in claims settlement.
For example, Lloyd’s International Trading Advice (LITA), a unit within Lloyd’s, provides insurance companies with information about government regulations in countries where it operates. LITA covers more than 200 geographies; the regulatory rules of each are within many unstructured documents. The manual process of responding by experts usually took several days per query. The LITA team trained a question-and-answer model using data from interactions, automating much of the process. Time to answer each query went down to less than an hour.
“The AI augments the role of the employees,” stated Craig Civil, Head of Data Innovation, R&D and Analytics at Lloyd’s. “It makes their job far more satisfying because we automate 80% of the work, and the 20% are truly interesting one-off questions that you do need an experienced team that can do the research and answer.
A new breed of customer service chatbots is also making impressions. Geico and insurtech startup Lemonade are using chatbots to settle claims, as well as handle many customer queries, freeing agents to handle more complex tasks.
NLP research continues to advance, as more complicated models become available. In the past year, OpenAI has released GPT-2 and Google has released Meena. Practical uses are likely to emerge from the advances, potentially allowing insurers to meet the needs of customers and save them money.