Artificial intelligence (AI) has become a part of most people’s daily lives — even yours. If you’re skeptical of that claim, then you might be underestimating the role AI has played in nearly every industry over the last couple of decades.
If you’ve ever asked Siri to tell you a joke, used auto-complete suggestions when writing an email, or taken a picture on a smartphone, you’ve engaged with a form of AI. One emerging AI tool that is especially relevant to you as an insurance agent is intelligent automation.
According to AI innovator IBM, intelligent automation (or cognitive automation) uses automation technologies to streamline and scale decision-making processes. More simply, it uses AI-powered software to learn how to automatically complete repetitive and routine tasks.
In this blog, we’ll focus on two areas where intelligent automation can benefit the insurance industry: underwriting and claims processing.
Using AI to automate, accelerate, and simplify the underwriting process is a major benefit that many tech companies claim their AI products can do. Intelligent automation is also touted as a way for insurance companies to get a leg-up on competition. But what does it look like on a practical level? Here's a simplified summary that’s free of Silicon Valley jargon.
The main message we saw from multiple data and technology companies was that AI is not meant to replace underwriters — it’s supposed to make them better.
Deloitte, a research and consulting firm, thinks insurance companies that embrace AI and automation will create exponential underwriters. An exponential underwriter will “leverage emerging tools, information, and skill sets to focus on higher-level challenges and become more strategic in defining the future of the company to enhance business performance and shareholder value.”
In other words, AI tools can help free underwriters from spending too much time gathering data to determine premiums and instead let them price risk more competitively and increase customer satisfaction.
How can intelligent automation help improve the underwriting process? Our research found these three common applications:
Information Gathering and Data Entry
An online policy application can gather an applicant's basic information, but tools like natural language processing, machine learning, and intelligent document processing can make the initial information-gathering phase a lot less time consuming.
Instead of requesting a client’s health record, paystub, or property value assessment and manually entering all the data into a system, an exponential underwriter could upload digital or physical copies of the required documents into a document capture tool. Then, using the AI tools mentioned above, the document capture service can analyze, organize, categorize, and input the required data automatically.
Another AI benefit for underwriting is access to more data for comprehensive risk assessment and fraud detection. Machine learning and language processing models can comb internal and external sources for relevant data, such as previous claim history, medical records, or a location’s natural disaster risk, to learn from the past and predict the risk profile of new submissions.
Data Assessment for Setting Premiums
After the initial data gathering and risk assessment steps are complete, intelligent automation tools can create data-driven dynamic pricing. Within the insurance industry, AI tools can analyze thousands of similar risk factors and create a custom pricing model for an individual applicant.
This benefits the insurance company by pricing risk more competitively and helps the customer by reducing the time it takes to offer policy pricing.
Claims processing is important for both the insurance company and the policyholder. According to an Ernst & Young market report, 87% of policyholders say the claim settlement experience impacts their decisions to remain with insurers.
On the business side, the AI tool creation company Scale found that 51% of surveyed insurers want to use AI to accelerate claims processing.
Like with underwriting, we’ll focus on three main ways that AI tools can improve claims processing. They include:
First Notice of Loss
If done manually, the first notice of loss (FNOL) can be a lengthy and time-consuming process. The two AI tools that are often suggested for this stage are chat bots and autonomous decision-making through internet of things (IoT) devices.
Chat bots powered by machine learning and language processing can effectively gather and organize all the required claim information from a policy holder through an app or website — all with little or no human interaction.
IoT, according to IBM, refers to a “network of physical devices, vehicles, appliances and other physical objects that are embedded with sensors, software and network connectivity that allows them to collect and share data.”
For example, with a claim involving a collision, an internet-connected car can record the time and place of impact, the areas of the vehicle affected, and details about the driving practices prior to the collision. The collected data is neutral, based on facts, and can serve as a basis for a claim that can be supplemented by later reporting.
Claim Assessment and Fraud Detection
Using neutral and fact-based data collected by IoT devices and organized by AI can also help improve claim accuracy and prevent fraud. But if a customer doesn’t have IoT devices in their car or home, computer vision is another option.
Computer vision uses AI to analyze images and videos. In this case, AI can reduce a claim adjuster’s investigation and inspection period by analyzing geospatial information collected from satellites, drones and customer videos or photographs.
One developer, ForMotiv, has developed multiple AI-powered tools that use predictive behavioral analytics to prevent insurance fraud. ForMotiv says that by analyzing patterns of behavior and interactions, predictive behavioral analytics can unveil anomalies that might signify fraudulent behavior.
Real-time behavior monitoring, for example, can detect sudden or unusual changes — like when a policyholder with no history of claims submits several high-value claims.
The prospect of utilizing new insurance technologies like AI-powered automation may be exciting for insurers and even some policy holders, but there are challenges to overcome. These include regulation compliance, privacy concerns, and the fact that sometimes people don’t want to “talk to robots.”
One of the problems with growing insurance technology adoption is that our current regulation system seems counterproductive to the insurance technology movement.
Thanks to the McCarran Ferguson Act, each state is responsible for its own insurance industry regulations. But insurance technology exists within the unbordered realm of the internet and the cloud. As insurance technology expands, it must find ways to maintain compliance within state regulatory schemes.
Automatic data collection through IoT devices also plays into the larger conversation over personal data and privacy within the tech industry. A policy holder might opt into allowing an insurer to access and use personal data for automation purposes, but that’s not a guarantee.
A Policygenius survey made it clear that insurers have a long way to go to establish trust in new AI-powered automation technology.
According to the survey:
Keeping the Human Touch
Deloitte interviewed chief claims officers (CCOs) from a dozen personal and commercial lines carriers and found that the desire to divert more claims to automated systems while maintaining a human touch is like walking a tightrope.
Instead of an either/or situation, Deloitte argues that insurers should invest in AI and proper training so human employees can effectively utilize the technology.
Additionally, the CCOs repeatedly highlighted the importance of personal engagement when clients need it, which should be a differentiator in an increasingly automated world. While an automated system might reject a claim because the particular loss isn’t covered, a human claims professional might use that opportunity to provide an additional value-added service.
AI is already part of your personal life, and it’ll likely become part of your professional life if it hasn’t already.
Nasdaq says 65% of insurance companies plan to invest $10 million or more into AI technologies in the next three years. If you want to dive deeper into the world of these emerging technologies, check out the sources below. As more insurers adopt AI tools in the coming years, it’ll be important to understand how to use them to your advantage!
Balasubramanian, R. et al. (2020, July 31). Rewriting the rules: Digital and AI-powered underwriting in life insurance. Retrieved from https://www.mckinsey.com/industries/financial-services/our-insights/rewriting-the-rules-digital-and-ai-powered-underwriting-in-life-insurance
Bassi, D. et al. (2017). Claims in a digital era. Retrieved from https://assets.ey.com/content/dam/ey-sites/ey-com/en_gl/topics/insurance/insurance-pdfs/EY-claims-in-a-digital-era.pdf
Burr, J. (2023, January 19). 9 ways we use AI in our products. Retrieved from https://blog.google/technology/ai/9-ways-we-use-ai-in-our-products/
Cline, M. & Kamalapurkar, K. (2021, October 12). Preserving the human touch in insurance claims transformations. Retrieved from https://www2.deloitte.com/us/en/insights/industry/financial-services/insurance-claims-transformation.html
Dilmegani, C. (2023, January 10). AI in Underwriting: Data-driven Insurance Operations in 2023. Retrieved from https://research.aimultiple.com/ai-underwriting/
Dilmegani, C. (2023, May 09). Intelligent Automation in Insurance 2023: Use Cases & Examples. Retrieved from https://research.aimultiple.com/intelligent-automation-in-insurance/
Dilmegani, C. (2023, October 12). Top 3 Claims Processing Automation Technologies in 2023. Retrieved from https://research.aimultiple.com/claims-processing-automation/
Dilmegani, C. (2023, October 12). Top 3 Ways AI Improves Insurance Claims Processing in 2023. Retrieved from https://research.aimultiple.com/insurance-claims-ai/
ForMotiv. (n.d.). Insurance Fraud Detection Software | Top 3 AI Insurance Fraud Solutions. Retrieved from https://formotiv.com/blog/insurance-fraud-solutions/
Gupta, M. (2023, April 17). Harnessing The Power Of AI In The Insurance Sector. Retrieved from https://www.forbes.com/sites/forbestechcouncil/2023/04/17/harnessing-the-power-of-ai-in-the-insurance-sector/
Howard, P. & Swartz, A. (2021, August 24). Policygenius survey: 83% say “no way” to AI handling their home & auto insurance claims. Retrieved from https://www.policygenius.com/homeowners-insurance/home-auto-technology-survey-2021/
IBM. (n.d.). What is intelligent automation? Retrieved from https://www.ibm.com/topics/intelligent-automation
Insurance Information Institute. (n.d.). McCarran-Ferguson Act. Retrieved from https://www.iii.org/publications/insurance-handbook/regulatory-and-financial-environment/mccarran-ferguson-act
Ladva, P. & Grasso, A. (2023, April 17). Benefits and use cases of AI in insurance. Retrieved from https://www.swissre.com/risk-knowledge/advancing-societal-benefits-digitalisation/opportunities-ai-insurance.html
Morris, C. (2023, July 19). AI Is Helping Insurance Companies With Underwriting and Due Diligence. Retrieved from https://www.nasdaq.com/articles/ai-is-helping-insurance-companies-with-underwriting-and-due-diligence
Scale. (2023, June 09). Guide to AI for Insurance. Retrieved from https://scale.com/guides/ai-for-insurance#introduction