By Benjamin Walker | Assistant Vice President – Cyber and Technology E&O Underwriting, Munich Re Specialty Insurance

One of the oldest and most-repeated sayings about insurance is, “It’s a people business.” With the advent and rapid development of artificial intelligence (“AI”), one might wonder if that saying will remain true. There are good reasons to believe insurance will always involve human intervention and human relationships – but AI and other technologies will play increasingly influential roles, particularly in underwriting.

Due to its ability to sort through large volumes of data, AI has the potential to support tasks such as assessment, analysis, and pricing of risk. These core functions of underwriting are repetitive actions that eat up a lot of time for insurance professionals. AI’s capability to alleviate that repetitive work could free up human underwriters to consider more complicated risks and to come to informed opinions more quickly and with increased accuracy. Let’s look a bit closer at what AI in underwriting might look like.

AI today is being used to classify risks and sort through submissions according to North American Industry Classification System (NAICS) codes or Standard Industrial Classification (SIC) codes. This is time-saving and helpful. It also gives human underwriters a second set of “eyes” able to flag information that may not have been captured or correctly classified. Deciding to accept a risk with missing or incorrect data could result in an insurer significantly underpricing that risk or inadvertently taking on risks outside the company’s appetite.

In cyber risk underwriting, for example, AI is widely used to conduct scans of public-facing information technology infrastructure. With this data, AI tools generate risk scores according to vulnerabilities uncovered by the scans. All of this happens quickly, presenting digestible insights that inform underwriting decisions. AI can analyze loss data for different threat vectors, determine a median risk score, and let underwriters apply different pricing scales to accounts above or below a certain score.

There is no question that AI applied properly may result in significant time savings for human underwriters. Hours saved enable humans to apply their expertise to other, more complex risk situations, and to develop new or enhanced coverage products. Eliminating redundancies and manual steps in underwriting will mean less administrative work, and faster quoting, binding, and issuance of policies. All of that will result in improved productivity.

Even though AI can be quite sophisticated, it’s important to remember that intelligence starts with human knowledge. AI tools are only as effective as the accuracy of their algorithms. Errors can occur anywhere humans work, and the same can be said of AI-driven automation. In addition, an expert underwriter understands the subtle differences in exposure that AI cannot. Therefore, it’s important to be cautiously progressive about implementing new technologies.

The combination of AI and human expertise offers the best resolution for risk challenges in cyber and technology E&O underwriting and other lines. An advantage that human underwriters have is the ability to talk with colleagues and help each other validate their perspectives on risk. An easily overlooked and important aspect is the value of personal touch in business relationships. As an example, responsiveness is important to policyholders and brokers. For these reasons, AI is unlikely to replace the human side of this people business and will serve as a valuable tool for the underwriter.