AI in insurance

AI in insurance is the use of artificial intelligence, automation and other advanced technologies to improve coverage and service delivery in the insurance industry.

Like other financial service industries, the insurance sector requires a large amount of data. This data helps carriers decide what insurance to give to which people and which premiums they should charge. Artificial intelligence can improve providers’ decision-making capabilities, driving increased care to their customers while improving their bottom lines.

The insurance industry has always made extensive use of data and algorithms, such as in the calculation of insurance premiums and processing personal and non-personal data in the underwriting process to assess risks and price insurance policies. But AI enhances those capabilities at increased scale and speed.

The rise of insurtech companies that use new technologies to serve customers can either provide services to legacy providers or challenge them for business.

AI-powered technologies can help organizations that are deploying insurance to individuals and companies alike. As such, insurance providers and other organizations within the insurance ecosystem should consider developing several AI-driven initiatives to realize the benefits of this powerful technology.

There are several AI applications that insurance providers can use to improve their operations.

• APIs
• Business Process Automation
• Generative AI
• Intelligent Automation
• Machine learning
• Natural Language Processing
• Optical character recognition

APIs enable software applications to communicate with each other to exchange information. APIs can connect the various types of organizations within the insurance ecosystem to collaborate. It can connect the carrier— third-party adjusting firms—and claimants to better share and access information. It is also responsible for continuing to grow the insurance industry. For example, the rise of insurtech companies like insurance intermediaries and aggregators can use APIs to connect with insurance carriers to show rates and offers to their customers.

Business process automation (BPA) automates complex and repetitive business processes in insurance. BPA can easily handle customer onboarding, claims processing, underwriting and other policy management services.

Generative AI (Gen AI) uses large language models (LLMs) can help insurance companies in multiple ways. Gen AI can help insurance employees streamline tasks such as answering customer service issues and analyzing documents or individual blocks of text. It can help customer service representatives better respond to customer issues. It also can help customer solve their own issues through the use of AI technologies such as chatbots and virtual assistants. For example, IBM helps companies optimizing processes that are used for handling large documents and blocks of text or images by using generative AI through its watsonX technology. IBM also created a chatbot for an insurance client that helped policyholders supply necessary documents and access the complete view of the coverages provided in their insurance package. Accordingly, 77% of industry executives said they needed to embrace gen AI quickly to match their competitors, according to a Institute for Business Value report.

Intelligent automation is a hallmark of any AI-driven workflows. It involves the use of automation technologies to streamline and scale decision-making across organizations. For example, an insurance provider can use intelligent automation to calculate payments, estimate rates and address compliance needs.

Machine learning (ML) uses data and algorithms to enable AI to imitate the way that humans learn, gradually improving its accuracy. Insurance providers can use ML technologies, such as deep learning, to analyze their customer data and power services that make product recommendations to prospects and customers.

Natural language processing (NLP) is a type of AI that uses machine learning to enable computers to understand and communicate with human language. Insurance companies can use NLP to parse the information customers supply to identify whether they can offer the right insurance and at what cost. For example, organizations providing healthcare insurance can ask prospects questions about their medical histories to better underwrite their insurance offering.

Optical character recognition (OCR), also known as text recognition, uses automated data extraction to quickly convert images of text into a machine-readable format. It is a crucial component of insurance companies’ approach to digitization, turning legacy assets into searchable digital content. Using OCR to digitize old forms and claims to put into a database can help them better understand the full history of their business and their service offerings.

AI solutions power several forms of insurance provider use cases.

• Claims management
• Code modernization
• Fraud detection
• Managing risk
• New product development
• Underwriting

AI tools can improve the claims handling process by expediting claims processing and settlements. Using Natural Language Processing (NLP), insurance carriers can read, interpret and process documents and images to decide whether to grant a claim or not.

By collecting large amounts of historical data, Discriminative AI can be used to make plausibility assessments and promote quality and uniformity in the adjusting process. Complimentarily, Generative AI will be able to help the adjustor summarize the data and generate a preliminary report.

Insurance companies, especially those founded decades ago, may use a mix of legacy technologies like Cobol, Assembler and PL1. IBM uses generative AI to help established insurance companies modernize their IT systems and create code that will work with existing technologies. Insurance carrier Sun Life used IBM Application Discovery and Delivery Intelligence (ADDI) solution to edit code, debug and initiate application discovery for impact analysis to its IBM zSystems® mainframe.

AI tools can improve the claims handling process by expediting claims processing and settlements. Using Natural Language Processing (NLP), insurance carriers can read, interpret and process documents and images to decide whether to grant a claim or not.

By collecting large amounts of historical data, Discriminative AI can be used to make plausibility assessments and promote quality and uniformity in the adjusting process. Complimentarily, Generative AI will be able to help the adjustor summarize the data and generate a preliminary report.

Fraud detection is the process of identifying suspicious activity that indicates criminal activity in insurance may be happening. The biggest example of potential fraud in insurance involves false insurance claims through the invention of an accident or embellishing what happened. Other examples include falsified medical records, using another person’s identity or other misreprensentations. AI can analyze the supplied data and compare it to historic data to better determine whether a claim is accurate or not.

AI-driven fraud detection software can monitors transactions, applications, APIs and user behavior to help organizations better prevent fraud or stop existing fraud in progress.

Risk management and risk assessment are key components of an insurance providers’ business strategy. To encourage profitability, carriers and reinsurers should understand the risk of each of their customers needing to make a claim. It holds true in every type of insurance. Using AI to analyze the large amounts of data an insurance company has from external events and that supplied by their customers may help them price their policies appropriately and attempt to minimize big surprises.

IBM is currently using property insurance underwriting and claims investigation to build foundation models in the IBM watsonx™ portfolio of AI products. The goal of the model is to improve the success and efficiency of risk evaluation and decision-making processes.

An Institute for Business Value study found that 60% of insurers predicted nontraditional products and services would soon generate as much revenue as existing, traditional products. Insurance carriers may dabble in differentiated risk environment, such as behavior-based insurance. They will need AI tools to better understand these environments and know how to accurately price their policies.

Underwriting is the process of deciding whether to offer an insurance policy to an applicant and pricing it appropriately. AI models can help companies improve their underwriting by analyzing the data supplied by customers. The carrier can choose to have AI decided whether to make the offer or not and to use the supplied data to price the policy.

The adoption of AI produces several benefits to carriers and other insurance industry organizations.

• Increased efficiency
• Improved cybersecurity
• Personalized customer experiences
• Predictive analytics
• Reduction of claims

Using AI tools like generative AI and machine learning helps organizations in the insurance industry better complete manual tasks like claims processing, new client signups and marketing and communications activities. Using AI to manage many of these tasks helps employees to tackle more important tasks like solving customers’ more difficult issues. AI also helps enhance workflows. IBM used its IBM Cloud Pak® for Business Automation offering to help transform the Swiss Re’s finance department’s quarter-end closings using low-code tools, reports and dashboards along with AI-driven analytics.

AI may help organizations better detect potential fraud of security issues. AI-driven cybersecurity may detect issues quicker and even potentially remedy an issue without human intervention. Given that insurance companies host important personal data, using AI can help avoid big reputational and regulatory problems.

AI helps organizations improve customer experience by marketing to specific groups using customized messages. It also enhances customer support by providing more powerful self-service customer service tools like chatbots and virtual assistants and equipping customer service representatives with more information through generative AI. The Institute for Business Value survey found insurers that used gen AI experienced a 14% higher retention rate and a 48% higher Net Promoter Score.

Insurance carriers can use AI in their data management procedures to improve their insights gather and analysis. They need to know what is likely to happen in the future that could impact their liabilities with relationship to existing policies. Using AI to derive predictive insights from existing data can help them shape strategy that will capitalize on today’s environment while avoid potential issues in the future.

AI can be used in the home for things like the Internet of Things (IoT) technology like carbon monoxide and smoke detectors to alert home owners in real time when a potential damaging incident is occurring. AI used in smart devices can also help reduce the risk of death in life insurance claims by catching potential life threatening situations or health issues.

Incorporating AI in insurance carries some potential risks, which companies should anticipate.

• Data quality
• Potential for discrimination
• Regulatory issues
• Skills gaps

Using AI alone can create some data issues. The technology is still improving, so it is possible for it to make mistakes, such as hallucinating data that is not there or making incorrect assumptions about a request. A miscalculation or insertion of phantom data could have a significant impact on strategic decisions that arise from that data. This reinforces the need to use employees to check the results that AI produced or use other types of checks and balances.

Since AI is trained on human data sets, the models can discriminate, either refusing to offer insurance to certain groups or overcharging premiums. There may be regulatory concerns if companies do not take adequate steps to curtail any potential issues with discrimination.As such, insurance companies likely should not use the more generic generative AI tools like ChatGPT and either work with companies that develop use-case specific tools, like IBM, or develop their own tools.

Insurance carriers must take steps to protect customer data. AI can help protect that data, but using it on external AI tools could be a violation of certain regulations. Insurance carriers should deeply investigate any AI tools they are considering use and seek the guidance of legal professionals before exposing any customer data to those technologies. The IBM watsonx.governance™ toolkit for AI governance helps insurance companies monitor and govern their entire AI lifecycle, minimizing risk and compliance issues.

Companies in the insurance industry may not have the right resources internally to take full advantage of AI. They may not have the right personnel or those current employees may lack the right skills. Like those in other industries, insurance companies should invest in AI upskilling and reskilling to prepare those employees for future jobs that include AI as a major component. And it should look to hire new employees that already have AI-related skills when it has job openings.