Software

How Artificial Intelligence Is Transforming the Insurance Underwriting Process

AI is helping to lower costs in underwriting, but it's not without its concerns.

J.P. Pressley is a contributing writer to BizTech magazine and an editor at Manifest.

Underwriters at Allianz and other insurance companies have long faced an intimidating challenge: the need to find specific answers across hundred-page documents before they can make an attractive offer. Often, this means the underwriting process takes days if not weeks per customer. But AI is dramatically changing this.

How Is AI Impacting Insurance Underwriting?

For years, insurers have used automated tools to accelerate labor-intensive underwriting tasks. But with generative artificial intelligence powering those tools, underwriters can make application decisions faster and finalize insurance policies tailored to applicants’ specific needs while aligning with insurer protocols.

The biggest competitive edge AI offers underwriters is “risk digitization,” by which “information from disparate sources and formats is automatically parsed, evaluated, and, through a risk taxonomy, mapped into a computer-readable format that machines can understand and triage,” as Google Cloud experts Stathis Onasoglou and Julius Von Davier write in this piece.

FIND OUT: How chatbots are revolutionizing financial services.

Once digitized, underwriters “receive what we call decision-ready risks with streamlined quotation workflows, accelerating decision-consistency, efficiency and turnaround time,” notes Sam Lewis, vice president of product for Cytora, in the same article.

Cytora’s risk digitization platform has helped insurers transform their workflows through Google Cloud’s Vertex AI platform to implement digital risk processing.

Other solutions have also been advantageous, even for those without extensive machine learning expertise. Amazon Web Services’ Amazon Bedrock simplifies the deployment, scaling, implementation and management of generative AI models for insurers, making it easier to address specific underwriting challenges, such as rule validation and decision justification.

In both of these instances, AI has proved an ideal underwriting assistant, capable of analyzing vast data sets (such as credit scores or social media activity), streamlining application reviews and improving fraud detection.

Click the banner below to learn how financial services can conquer compliance and risk management.

bt-priorities-finance-animated-2024-desktop

bt-priorities-finance-animated-2024-mobile

The Benefits of AI in Underwriting

AI has reduced the average underwriting decision time of three to five days to 12.4 minutes for standard policies, while maintaining a 99.3% accuracy rate in risk assessment, according to a 2025 technical analysis. For complex policies, AI has helped reduce underwriting processing times by 31% while improving risk assessment accuracy by 43%.

This explains why over 380 companies (including technology vendors and established insurance companies) now rely on AI-based underwriting as a second set of eyes to catch details that may otherwise be missed. Allianz moved forward with their new BRIAN solution, a generative AI underwriter guidance tool.

“AI has the ability to discern patterns in ways and in data sets where humans simply cannot, or they simply just don’t have the capacity to look at ginormous data sets and tease out various patterns,” Doug McElhaney, partner at McKinsey, tells BizTech. “That’s what AI can do very successfully.”

The Challenge of AI in Underwriting

Despite its benefits, using AI for underwriting poses some significant ethical considerations, particularly when it comes to unintentional bias.

If an AI system is trained with biased data, intentional or not, the insights will be inherently biased. This makes it essential that IT leaders “interrogate” their data to identify and eliminate potential biases early.

“Most traditional insurers underwrite life insurance based on five criteria: age, gender, ZIP code, whether you are a smoker or a nonsmoker, and marital status,” Bryan Simms, co-founder and president of Mammoth Life, explains in an industry article. “That can lead to, and has led to, very narrow definitions of acceptable risk for insurance carriers.”

That’s why this minority-owned company has its AI models pull from hundreds of publicly available data points that most traditional insurers don’t, giving marginalized groups a fairer risk assessment.

Other insurers incorporating AI in underwriting would do well to evaluate their processes to ensure they also use AI to combat bias, not amplify it. As Russell Page, CTO of Hagerty, a leading specialty provider for the global collector vehicle market, tells BizTech, “What you never want to use AI for is something that marginalizes or creates harm to any one group.”

UP NEXT: The biggest tech trends financial services will follow in 2025.

PeopleImages/Getty Images