AI can cut costs and personalize experiences for insurers. But success requires a modern data architecture and intelligent data management.

The insurance industry has traditionally been known for its extensive paperwork, lengthy processes and arduous risk assessment procedures. But the time for tradition is over.

Severe weather events and continued inflation are wreaking havoc on insurers’ top and bottom lines. The US saw $34 billion in insured losses in the first half of 2023 due to damaging storms, according to reinsurance giant Swiss Re.

Meanwhile, inflation has played an outsize role in the 5% to 7.5% increase in P&C claims payouts in 2022 across five key markets globally, Swiss Re says, with additional increases expected this year.

Amid these rising costs, customers increasingly demand easier to use and more personalized insurance services. As a result, insurers can no longer let themselves be dragged down by expensive and inefficient processes.

Insurers are increasingly turning to artificial intelligence to meet the speed, efficiency and personalization they need while also lowering their cost threshold. In addition to reducing processing times from days to minutes or even seconds, smart use of AI can help insurers provide personalized and even proactive service to customers. An example might be offering a small commercial business owner a personalized quote, triggered by a real-time event like the opening of a new outlet.

But enabling these capabilities requires ongoing investments in data quality, data management and a data architecture that enables the AI system to deliver on its insights, personalization and efficiency capabilities.

The potential of AI

With the proper use of AI, insurers can not only reduce costs by automating processes; they can also improve their processes to fit today’s needs. For example, AI can process information in various formats, customize pricing models and streamline underwriting and claims processing.

In underwriting, we’ve worked with P&C clients to use natural language processing to identify the most important pieces of information in their property inspection reports and performance statements. This has led to a 10X reduction in underwriting time per case and an expected 25% increase in case acceptance due to more efficient underwriting.

Our life and health insurance clients have used machine learning-driven prediction to achieve 98% precision in identifying which applicants were actually non-smokers vs. those who claimed to be so. By doing so, they were able to price policies at a rate that reflected actual health status.

A global P&C insurer used AI to create an intelligent assistant that reduced claims processing from days to minutes. Now, when calls come in, the system can detect the caller’s level of frustration or satisfaction. Service agents receive real-time quality assessments, as well as a personality profile of the caller and conversation cues to enhance support. As a result, supervisors now spend 40% less time auditing customer calls.

The system also generates voice call transcripts with high enough accuracy (85%-plus) to produce deeper insights about the customer’s experience and needs. And with new self-service capabilities, as many as 60% of customers have been able to file their first notice of loss themselves, increasing their satisfaction while reducing cost and delays.

Yet another insurance client used advanced analytics to create a fraud detection engine that identifies fraudulent claims that had previously gone undetected. By doing so, they discovered a very small set of claims (the top 0.26%) with a very large fraud hit rate (40%). By focusing on these claims, the insurer reduced costs and now has a faster, easier way to conduct its fraudulent claims investigation.

Now, with generative AI, insurers can create differentiated customer and employee experiences, reimagine business processes and speed AI solutions to market at enterprise scale.

Preparing for AI

Using AI effectively requires large quantities of reliable, personalized, secure, real-time, easily accessible data. To make that data available, insurers need both a modern cloud data infrastructure and the intelligent tools and techniques to manage the data.

Data quality and accessibility are essential for enabling AI to make actionable business recommendations in real-time. For example, models trained on inaccurate data cannot pinpoint the exact right time for an agent to reach out to a customer or correctly customize an offer for them. In fact, creating a more data-driven decision-making environment is increasingly a top concern for CEOs.

We advise our insurance clients to consider the following to prepare for AI:

• Modern data architectures like data mesh or data fabric, which distribute data governance among the teams that own the data, improve data management and security, and optimize access to distributed data that is curated and orchestrated for self-service delivery.
  
  
• Intelligent modernization tools, such as Cognizant’s Data and Intelligence Toolkit, Intelligent Migration Studio, to manage—cost effectively and at scale—the discovery, migration, security and integration of the data needed for AI applications.
  
  
• Data quality frameworks, such as Cognizant’s Data and Intelligence Toolkit, DataIQ, that can evaluate the relevance and intelligence of their data assets, enabling them to focus their data engineering efforts on the data that will help them deliver the desired business outcomes.
  
  
• Platforms that accelerate the adoption of generative AI technology, such as Cognizant’s Neuro®️ AI, in a flexible, secure, scalable and responsible way. Such platforms should include a library of reusable generative AI models and agents, development tools and control components, as well as versioning and auditing capabilities.

The future of insurance with AI

The era of costly and inefficient insurance processes is over. With a modernized data foundation, insurers can optimize their use of AI to meet the speed, efficiency and personalization requirements of today.