How insurers can overcome 4 key challenges when ramping up Gen AI

We are only beginning to scratch the surface with generative AI. Last year, many companies and individuals were exploring its potential. But now companies are looking for more advanced uses within their organizations. In insurance, this means exploring generative AI applications across claims, underwriting, and other processes. Each of these changes requires an intense focus on testing, workflow transformation, customization, and integration. 

Investment in AI will impact the entire insurance industry in many ways. But with this new technology comes a new set of challenges insurers need to be prepared to address. Here are the four biggest challenges insurance organizations will encounter as they ramp up their generative AI usage, and how they can overcome them to provide value to their organization. 

1. Building trust in computer generated answers. Over time, insurance organizations will become more and more comfortable with computers making choices. These computer-led decisions will include everything from the level and type of coverage recommended to claims decisions. Today, while many insurers are using machine learning to automate tasks in claims and underwriting, generative AI requires them to take the next step and allow the technology to make its own inferences. To become comfortable and gain trust in its findings, insurance organizations will need to invest in transparent AI systems that undergo rigorous testing. Testing should include a variety of real-life scenarios, as well as adversarial testing with false information to check performance. And perhaps most importantly, communicating the results is vital. It's also critical for insurers to put governance frameworks in place to address bias, safeguard data and ensure privacy. 

2. Updating workflows to match current AI capabilities: Most insurance processes have been known to be slow to change. In fact, the questions insurers ask clients to quote a policy haven't been significantly changed in decades. While generative AI can make workflows more efficient, it will require employees to adapt to the changes. For example, extracting and interpreting data from multi-page documents is time consuming for underwriters and claims managers and is prone to human error. Using large language model solutions, information can be automatically extracted from documents and used to complete forms. To help employees embrace solutions, insurers should put adequate training in place and have resources available to troubleshoot issues. 

3. Shifting from clusters of risk to tailored policies: Insurance models have historically considered clusters of risks in determining rates and coverage. But AI changes the game, giving insurers the ability to quickly analyze large values of data and look at risks individually. For example, with auto insurance, providers usually rely on personal details, driving history and vehicle information, but now they can factor in telematics data, location and geospatial data, social media, and sensor data to create policies specifically tailored to individual drivers. 

This type of customized view can even give insurers new opportunities from clients that they may have previously rejected, consider new types of business, and identify new market approaches. As generative AI becomes more widely used, insurers will be prepared to offer policies based on individual, real-time data instead of relying on historical clusters of data.

4. Driving usage by integrating data sources: An AI solution is only effective if employees use it. In order to make solutions work for team members, all various data sources should be integrated into the platforms they use. This will enable users to leverage AI for informed—and ultimately, better—decision making. A great option is for insurance organizations to create sandbox environments for generative AI. There, employees can experiment, explore AI capabilities, identify valuable use cases, and adapt the technology to their needs. Insurers can also build multidisciplinary teams of business experts, IT specialists, and data scientists to guide the integration of AI into organizational processes. 

We're just at the beginning of generative AI's influence. More advanced applications will enable insurers to streamline processes, improve the customer experience and grow their business. However, with new advancements come new challenges. Robust testing, new workflows, and enabling employees to experiment and explore AI capabilities, will enable insurance organizations to bypass the challenges and capitalize on all of the benefits this technology has to offer.