Can Produce, Can Underwrite: AI in Underwriting at ACORD

This ACORD Data & Analytics track panel discussed the promise of artificial intelligence for enabling ‘human augmentation’ for underwriters and brokers.

(Panelists from left to right: Mike Fitzgerald, Frank Neugebauer, Ram Chatty and Rajesh Narayan.)

One of the standing jokes in the insurance industry is that CPCU stands for “can’t produce, can’t underwrite,” but a session at ACORD2017 gave the saying a new twist. The discussion, “Can Produce, Can Underwrite (CPCU),” held Oct. 26 as part of the show’s Data & Analytics track, explored the potential of “human augmentation”—the use of artificial intelligence (AI)—to improve the work of underwriters/producers, especially in the case of less experienced professionals.

Mike Fitzgerald, Senior Analyst, Celent.

“Much of the discussion is around the underwriter and broker personas,” clarified panelist Mike Fitzgerald, Senior Analyst, Celent (Boston). “We are not changing the underwriter with AI, but we are changing underwriting to take advantage of advances in technology.”

Fitzgerald and other panelists established that AI has advanced at a rapid rate and as a consequence has become more realistic—and affordable—technology for insurers to deploy. The term AI, according to the panelists, provides a variety of capabilities and benefits useful to insurers, including recognition of complex patterns in risk identification and the matching of for those insights to products; the capacity to offer underwriters and brokers real-time information—such as identifying the right product, market or risk—based on complex rules and statistical analysis; and enabling related technologies, such as computational AI and computer vision.

Force Multiplier for Core Systems

For example, panelist Ram Chatty, VP, Digital, Genpact (Hamilton, Bermuda), commented on the ability of AI to help process unstructured data in a way that minimizes the administrative burden of professionals and enables them to focus on high-value tasks. Chatty introduced a use case that, he glossed, “combined computational linguistics to read unstructured emails and attachments, triage the submission using internal and external data sources to understand [their] match to risk appetite, and also detect the propensity to bind from past producer and underwriter behavior.”

Rajesh Narayan, AVP, Insurance, Genpact.

AI can also act as a kind of force-multiplier to core insurance systems, noted Frank Neugebauer, VP, Digital Business and Insurance CTO, Genpact. “The value you get from a core system modernization is only incremental over the mainframe,” he said. “But when you couple it with digital interventions such as AI and the use of analytics, you can get a much higher business value from these initiatives.”

The discussion ranged into the use of drones, and the data they produce, in catastrophe events. Drones can give quick insight into the magnitude of the event overall, while also providing detailed information through their ability to access structures such as rooftops without any risk to claims professionals, the panelists affirmed. They also explored computer vision—technological capabilities for acquiring, processing and analyzing digital images. While this AI-related area is still largely experimental, it shows promise for automating aspects of the claims adjustment process, for example, analyzing a photograph to determine whether a damaged automobile should be treated as a total loss, the panel affirmed.

AI, The New Kid on The Block

Frank Neugebauer, VP, Digital Business and Insurance CTO, Genpact.

AI itself remains in the early stages and presents difficulties for insurers considering adopting the technology, according to the panelists. For example, legacy systems don’t lend themselves easily to AI in the absence of strong integration capabilities. Also, insurance professionals may either resist deploying AI out of discomfort or because in their incremental way of thinking they struggle to understand the kinds of leap-frogging that AI can enable in insurance processing.

For now, AI is still “the new kid on the block,” quipped Rajesh Narayan, Assistant VP in Genpact’s Insurance business, where he specializes in underwriting. “It will take a village—the insurance industry—to help nurture AI’s potential,” he remarked. “However, if done right, this technology can change the industry for the better, helping to personalize insurance, create new markets, and grow the book of business profitably.”

Three Approaches to Artificial Intelligence Value

Anthony R. O’Donnell // Anthony O'Donnell is Executive Editor of Insurance Innovation Reporter. For nearly two decades, he has been an observer and commentator on the use of information technology in the insurance industry, following industry trends and writing about the use of IT across all sectors of the insurance industry. He can be reached at or (503) 936-2803.

Leave a Comment