Putting a premium on AI use to transform insurance

Artificial Intelligence (AI) is like a swiss knife—versatile, powerful, and incredibly useful. The ideal assistant for anyone and everyone. The beauty of AI lies in its ability to shield users from its complexities, while delivering human-like abilities in carrying out tasks. For enterprises, many of the backend complexities are easily handled by Amazon Web Services (AWS) offering a smooth, enhanced experience.

Expanding the boundaries of insurance with AI

As technology advances, so too will the capabilities of AI - enhancing customer experiences across industries, such as insurance. At ICICI Prudential Life Insurance Company, Karthik Kanagaraj is spearheading this transformation. “With advances in AI, our ability to integrate text, speech, vision, and combine it with medical reports or wearable data, could soon help underwriters make informed decisions,” shared Karthik, Chief Data Science and Analytics, MD & CEO Office, ICICI Prudential. This is transformative for the industry, as underwriters carry out the vital task of assessing the risk of insuring individuals, companies, assets and more to determine the conditions of the insurance policy.

Interestingly, Karthik, who possesses chemical engineering degree from Anna University and MBA from IIM Bangalore, has worn multiple hats in his over two decades career in the insurance sector. He started in sales, moved on to strategy, then finance and is now spearheading technological advances in insurance. After college, his first job was at Colgate. Later in 2004 moved to ICICI Prudential, and has been there since then.

“My introduction to AI, data science, and machine learning happened in the last six years. The great thing is that at ICICI you can have different roles in your career,” said Karthik. When he was introduced to AI, he immediately saw opportunities where it could help the company, by converting data into intelligence. This was especially true in areas such as customer experience, claims servicing, risk management, identification of risky profiles and persistency (a metric used by insurance companies to determine the number of policyholders who will continue to pay premium).

“AI helps indicate to us which customers may not end up paying the scheduled insurance premium. It helps us reach out to them well in advance and share the benefits of insurance. If they don’t pay, they lose out on the benefits. So, AI helps us understand which customer will need guidance,” said Karthik. Another area where AI has helped is in product recommendation. Instead of shooting in the dark, marketing teams are better informed, with AI helping match potential customer profiles to products. With AI, various teams from underwriters, to marketing, customer support and others have seen productivity gains in their work.

Managing Scale Seamlessly

Interestingly, customers are not exposed to any of the complexities that go behind AI models. If an ICICI Prudential customer requires additional support from a customer service perspective, that knowledge is passed on by AI to the executive. “The overall acceptance for AI tools and AI interventions is quite good,” said Karthik. That comfort level is great particularly as ICICI Prudential deals with millions of policies, adding new customers, processing claims and more.

To manage that scale and intelligence in the system, AWS infrastructure is key. “Their infrastructure is our core. We use AWS SageMaker, our primary platform for model development and analyzing data. While AWS Bedrock is helping us set up generative AI use cases,” said Karthik. Data resides on Amazon S3 (Amazon Simple Storage Service), a highly scalable, durable, and secure object storage service designed for industry-leading scalability, data availability, security, and performance. AWS SageMaker, on the other hand, is a cloud-based machine learning (ML) platform that facilitates the creation, training, and deployment of ML models. SageMaker can also deploy ML models on embedded systems and edge devices. Bedrock is a fully managed service that offers a choice of Foundation Models (FMs) from leading AI companies like Anthropic, AI21 Labs, Meta, Stability AI and Amazon, through a single API, along with a broad set of capabilities that can be used to build generative AI applications. ICICI Prudential also uses AWS EC2 (Elastic Compute Cloud), which allows the company to run applications on the public cloud.

AWS solutions offer performance with scale. Karthik said, “the scaling ability is already built in. For our processes, where there is a new model deployment, we don’t have to worry about thousands of policies being logged in at the same time. All that scaling up happens automatically. It removes the infrastructure worry from our minds and we can focus on our core business.” AWS, he added, also makes customer experience uniform and well regulated.

Outside of his work in insurance, Karthik enjoys watching Hollywood movies and reading books when he gets time. James Clear’s Atomic Habits is among the books he has recently read and enjoyed. The book emphasizes the importance of small changes that can create great results over time.

Increasing Human-Machine Interactions

In the future, Karthik sees AI assisting insurance companies in more areas, such as reducing claims processing times or even detecting fraudulent claims. “AI in the future can also combine structured data with unstructured data and enable us to provide better customization,” said Karthik.

“Our daily habits are also changing with AI. We are already interacting with a lot of intelligence built into machines around us, like Amazon Alexa. Machines will also become more comfortable with multiple languages, similar to how people often speak by mixing both Hindi and English,” noted Karthik.

In this constantly changing tech landscape, Karthik and his team are always a step ahead thanks to a culture of training, development and learning within ICICI Prudential. They also get support from AWS via regular updates and periodic symposiums, which helps Karthik know where they are in terms of new technologies and benchmarks in the industry.

In the coming years, AI will have a wider impact across users and enterprises. Experts in the field point to Singularity (when machines will match human intelligence) happening in a couple of decades. And that will have an impact in multiple ways. “AI is going to remove a lot of repetitive and mundane tasks. Some of these things may be handled entirely through automation, which will allow people to focus on more creative endeavours,” said Karthik.

The industry is already registering these changes, as all new applications are assessed by AI, which determines which policyholders should receive reminders or communication.“Overall, in every interaction there’s increasingly an AI element in the back, like an invisible hand, which is built into customer journeys, granting them a better experience,” added Karthik.