Everyonetalks about artificial intelligence (AI) in the insurance industry and theenormous changes we will see in the future. However, AI is already used in insurance agencies in small ways.

Some insurance agencies have turned to AI software in their accounting departments. Off-the-shelf software developed for other businesses is readily available for use in insurance agencies. We also see chatbots added to agency websites.

Some chatbots can be as simple as a pop-up that appears to be a live person offering to direct consumers to a page on the website. Or, more complex bots collect data provided by the visitor for application purposes. In some cases, once the bot has exhausted the website's information, it can transfer the visitor to a live person along with a partially completed application so the human can complete the sale. This results in less time spent by agency employees talking to uninterested prospects while also maintaining 24-hour communication with consumers.

AI adoption as an alternative to third-party vendors is one of the largest growth areas. For the past decade, third-party vendors have leased employees from other countries to insurance agencies to handle routine tasks. These tasks do not require licensing or a personal presence within the U.S., and this is the sweet spot for using AI. More vendors are coming out with insurance-related products to handle many of the same tasks handled by leased employees.

From an errors & omissions perspective, AI can review an application and binder, checking the policy issued for language that might contradict the expectations. Policy language is flagged and provided to the agency to review to make sure the policy meets expectations. This AI function protects against the many E&O claims that occur when a policy exclusion or sublimit alters the coverage the agency intended to procure for its customer.

AI does not make the final call as to what is and is not covered. A person at the agency does that. However, it saves countless hours of tedious work reviewing applications, binders and policy language. Furthermore, because the task is being done by AI, it can be completed in minutes instead of hours.

While this software is helpful in the admitted market, it pays for itself in the excess & surplus lines market, where carriers can change coverage on a daily basis and without the blessing of a state's department of insurance. In a hard market, many admitted markets are no longer accessible and more business is being placed in the E&S market, making this a good time to adopt an outside source that specializes in policy review.

Still, AI is is only as good as the parameters set for it. Recently, an agency's claim stemmed from a carrier in a tower of insurance that did not include defense coverage in its layer. When a claim reached that layer of the tower, the insured was forced to pay its defense, while the carrier controlled indemnity. The carrier refused to settle and forced the claim to trial. While the carrier eventually settled the claim, the situation left the insured with a seven-figure defense invoice. The insured then sued its insurance agent for failing to ensure that all of the carriers within the tower provided defense coverage.

What was the agency doing to review policies and make sure they meet expectations? They were using an AI vendor. In this claim, the root of the problem was in the contract between the AI vendor and the agency. The contract contained a list of items the AI software would review—and matching coverages in layers of a tower was not included in that list.

When implementing AI in your agency, here are three tips to avoid an E&O claim:

1) Understand the product and what it can do.

2) Test the product before purchasing it to make sure it works as promised.

3) Continue to do things the way you have always done them while also using the new AI product.

Overall, continue to have someone review the policies and compare their results with the results of an AI policy review product. By doing so, you can tweak the AI product until it can at least match—if not surpass—the performance of your current human process. Once you have reached that level, you can potentially wind down the manual process and rely more on the AI product.

While this approach creates redundancy and extra costs in the process for a period of time, it is a necessary step to make sure you are getting what you are paying for. It will also reduce your chances of experiencing an AI-related E&O claim. Agencies are also encouraged to seek attorney advice to ensure the AI tool being used complies with applicable laws and regulations.

Jim Redeker is senior underwriter at Swiss Re Corporate Solutions and works out of the office in Kansas City, Missouri. Insurance products underwritten by Swiss Re Corporate Solutions America Insurance Corporation, Kansas City, Missouri, a member of Swiss Re Corporate Solutions.

This article is intended to be used for general informational purposes only and is not to be relied upon or used for any particular purpose. Swiss Re shall not be held responsible in any way for, and specifically disclaims any liability arising out of or in any way connected to, reliance on or use of any of the information contained or referenced in this article. The information contained or referenced in this article is not intended to constitute and should not be considered legal, accounting or professional advice, nor shall it serve as a substitute for the recipient obtaining such advice. The views expressed in this article do not necessarily represent the views of Swiss Re and/or its subsidiaries and/or management and/or shareholders.