Is Artificial Intelligence in Insurance Showing Tangible Results?

By March 27, 2019 February 17th, 2020 Blog
AI and ML in insurance

Incorporation of artificial intelligence (AI) applications in insurance are disrupting traditional insurance practices. AI use cases are becoming more common where insurance companies have achieved better customer satisfaction. Automation has helped reduce costs for insurers, enabling them to provide better coverage for lower premiums, resulting in enhanced customer loyalty and retention.

There is still a long way to go before AI completely penetrates the insurance industry, and companies pass on the benefit accrued to their target customers. Early adopters are however reaping the maximum rewards. Some areas where the insurance industry has implemented AI to yield tangible results are:

  • Usage-based insurance
  • On-demand insurance
  • Peer-To-Peer (P2P) insurance

Usage-based insurance (UBI)

UBI, better known as pay as you drive (PAYD), is a tailor-made vehicle insurance policy based on the type of vehicle, distance, run type, and driving behavior of the individual.

Data acquired using telematics and mobile apps are used to determine the premium for each individual. Telematics devices connected to automobiles capture data from the onboarding kit, providing vehicle information. This, along with driving behavior and combined with GPS information, allows the insurance companies to determine customer behavior. Mobile apps that capture the customer behavior (speeding, braking, acceleration, and distraction) have been developed by combining the GPS information and phone usage while driving.

Differentiation based on individual driving behavior allows the insurance companies to offer lesser premiums for safer drivers. This leads to greater customer satisfaction and an enhanced customer base. This is indicated from a survey conducted by JD Power & Associates in 2016 which found that “…UBI participants provided more positive recommendations and more often indicated that these recommendations resulted in a friend, relative or colleague purchasing from their insurer compared with those customers who did not use a UBI program.”

Earlier, insurance companies used to gather historical claims data and geographic data and run analytics to predict the future. Presently, telematics devices, sensors, and smart phones are the devices that provide insurance companies with the data they need to better assess the customer as ‘risky’ or ‘safe’. Based on the driving data captured from these devices, plus the legacy data they have, a large number of behavioral models and algorithms are developed. These models give almost accurate predictions based on the datasets provided. With increase in data, the model enters a self-learning phase, i.e. it matures and provides more accurate results. In future, companies will be moving their pricing models from likely predicted pricing to actual driving behavior of the customer in real time.

On-demand insurance or pay as you go insurance

On-demand insurance policies can be bought with the click of a button, whenever and wherever required, without directly interacting with a broker or a company representative. These policies include coverage for the on-demand economy (such as Uber drivers); they provide flexibility of term and pricing. Applications like Chabot have enabled the insurance companies to automate the policy buying experience of the customers. Based on the information provided by the customer, Chabot can be designed to pull social and geographic data to provide better coverages with good premiums in shorter duration. Better and faster interaction can lead to more business. These apps can be used to provide on demand or pay as you go policies. Companies like UBER, Lyft, Turo, Vehicle Rental agencies, etc., are already partnering with insurance companies to provide the drivers with on demand insurance for a specific period.

Automated platforms provide better efficiency with reduction in issuing time of policy. The various processes involved, from identifying the individual, to authenticating the credentials, fixing a quote, and binding a policy are all automated to provide better customer experience. According to a survey by Accenture, 74% of customers are willing to interact with computer generated systems for insurance advice.2 This personalized experience is more attractive because of the convenience and practically zero response time.

Peer-To-Peer (P2P) insurance

Insurers are designing models to speed up claim processing while reducing frauds to make way for a better customer experience. The obvious choice to achieve faster processing speeds in real time is by creating better AI models. This can be achieved by combining social media data, geographic data, legacy claim data, data from telematics devices, and GPS data from cell phones to assess the claims in real time.

Faster claims processing and better coverages lead to better customer satisfaction and increased sales. However, to achieve this, regular analytics will not be sufficient. Self-learning AI models that can read huge data sets from various sources in real time and generate accurate results will lead to better business outcomes.


The adoption of AI in the insurance industry promises to bring about the much needed change from traditional insurance practices to a more customer-and business-centric approach. Self-learning AI models capable of generating accurate predictions in real time promise to lead to better policy coverage, premium estimation and settlement.

With an increase in customer willingness to shift to smart technologies, greater adoption of AI in insurance promises tangible growth in business with reduced costs and greater customer satisfaction.


Chandrasekhar Ganta

Chandrasekhar Ganta

Chandrasekhar Ganta is an Associate Technical Architect at InsurAnalytics. Read More Posts

Leave a Reply