In early 2023, artificial intelligence, “AI,” became a hot topic of discussion. AI, sometimes called machine learning, is made possible by having sufficient data to train the models that AI is based on.
However, AI is not new–it has actually been around since the 1950s. Prior to 1949, computers lacked the programming for intelligence and could only simply store and execute demands.
Additionally, computing was extremely expensive and due to the cost, the use of computers was limited to large universities and technology companies. According to an article by Harvard University, it wasn’t until five years later that the proof of concept was actualized by the creation of Logic Theorist, a program designed to mimic the problem solving skills of a human.
Since the 1950s the development of AI has had its successes and setbacks, however as advancements were made, computers became less costly, quicker and more available.
The 1970s and 1980s welcomed a promising era of AI models that were making strides in development of technology that achieved the necessary intelligence to transcribe and translate spoken language as well as attain high throughput data processing, according to Harvard.
Critics of the progress of AI have remarked on the lengthy timeline of its development. Investopedia reports that AI models were not to blame. It was actually the computer storage capacity. Moore’s Law was inspired by Gordon Moore, the co-founder and former CEO of Intel. The law suggests that the number of transistors in microchips will double every two years, therefore making computing faster, more compact and efficient for years to come.
Time has been our friend in the realm of technology as AI has become more advanced over the past decades. Currently in our “big data” era, AI now helps humans to process large amounts of data that no one person could handle.
AI has proved its worth in many fields including entertainment, banking and marketing. Insurance has been embracing AI for some time now, but is now faced with the decision to implement the technology in other roles where it has not been utilized before.
More and more insurers are turning to AI to increase efficiencies, reduce costs, and lower fraud. DataIQ estimated that the use of AI has resulted in cost savings of over $1.3 billion for insurers providing auto, life, property and health coverages.
There are many uses for AI in the insurance industry, including underwriting, claims processing, data analytics, fraud detection and regulatory compliance. With the use of AI, underwriting processes can be streamlined, and incidences of human error may be reduced.
Automation of repetitive tasks leaves more time for underwriters to deal with complex problems and judgement calls. Predictive analytics can help carriers estimate potential losses with greater accuracy, including severity and repair costs. Real time data can be used in the quoting process, fine tuning the rates to the actual exposure. For example, use of historical claims data can aid in underwriting evaluation.
The claims process, vitally important to any insurance relationship, can be greatly impacted by AI. A study by Ernst & Young found that 87 percent of policyholder respondents believe that the claims’ handling experience has an impact on whether they stay with their insurer.
An inefficient claims process increases both insurer costs and customer dissatisfaction. AI can be used to analyze claim photos, to interact with clients/claimants and issue payments.
Fraud is a continual concern for insurers and AI can be used to uncover scams and fraudulent claims. Large data sets can be analyzed to detect patterns that humans might overlook.
A generative AI model could be used to gather the data set utilizing past claims records, social media activity and geolocation data.
The model would learn from the data collected, by identifying patterns and correlations. When the model is fully trained it would then assess the data and look to identify the likelihood of risks associated with fraud.
A human analyst would then assess claims flagged as potentially fraudulent to determine whether further investigation was warranted and/or approve or deny the claim.
Customer retention is the goal of all insurers and retaining existing clients is a cost-effective strategy. Each year a certain percentage of customers “shop” their accounts with other insurance companies. Some of them will leave no matter what. Churning is the percentage of customers who leave a business during a certain period of time.
According to TechSee, a 2019 study showed that 50 percent of customers shop their insurance policies each year. Good customer service practices like prompt responses to inquiries and properly trained employees help add to the customer experience and contributing to higher retainment rates.
AI can help carriers pinpoint those customers who shop, but even more importantly, those who can be convinced to stay.
In addition, a large percentage of call center contacts are made during non-business hours. Using AI, questions can be answered 24/7.
Virtual consultants can answer routine inquiries on coverage and claims processing, and even assist with password resets.
The National Association of Insurance Commissioners (NAIC) is very involved in analyzing the use and impact of AI in insurance. The NAIC has published guiding principles which address the use of AI by “AI actors,” which includes insurers, rating and data providers and advisory organizations.
The NAIC recommends that AI actors be guided by these principles: fair and ethical, accountable, compliant, transparent, safe, secure and robust.
AI actors should be sure to follow business best practices in the same manner as they would without AI.
In New York, the Department of Financial Services (DFS) has been looking at different ways to incorporate AI into regulatory processes to enhance effectiveness.
In an article published by Bloomberg, they referenced DFS Superintendent Adrienne Harris’s interest in purchasing a supercomputer to run AI systems. Superintendent Harris stated that the Department’s intention would be to use data driven approaches to leverage data analytics and enhance their ability to predict and respond to events in the marketplace.
Some people fear it, some people cheer it, but the use of artificial intelligence is here to stay. Understanding the many ways that AI can improve our work processes will be key to the future success of insurers. The use of AI technology should be a complement to the work of underwriters and claims analysts and should not be thought of as a replacement.
Like all solutions for insurers, proper education, training and implementation is necessary. Utilization of AI in insurance practices is not likely to be a passing fad, but the industry will need to identify the smartest ways to implement it—improving the experience for both insurers and their customers.
This article was previously published in NYIA’s Your NY Connection Magazine and is provided courtesy of MSO®, Inc. (The Mutual Service Office, Inc.) for noncommercial use only. For any other licensing requests or permissions, please contact firstname.lastname@example.org.
© MSO®, Inc. 2023.
MSO provides advisory services for all property and casualty lines except workers compensation. This includes customized forms and manuals for insurers, MGA’s, and agents / brokers.
Additional information is available at www.msonet.com.
This article is for educational and discussion purposes only and it is not insurance or legal advice and should not be relied upon when making insurance or legal decisions. Nothing herein shall be construed to constitute a legal or underwriting opinion. Nothing herein shall be construed as offering any political, social, or public policy opinion by the author or MSO. Neither the author nor MSO are responsible for errors in, or the accuracy or currentness of, the article.