Generative AI holds key to future of insurance innovation

By Cynthia Ezekwe.

The potential of generative AI to revolutionize the insurance industry is immense. By harnessing its ability to create new data based on existing data, insurance companies could speed up the process of pricing policies and processing claims, and even develop new products to better meet customer needs.

Generative AI is a branch of artificial intelligence that creates new content based on existing data. Unlike other types of AI that analyze or predict outcomes, generative AI can be used to generate images, text, audio, and even video. It does this by using algorithms to analyze the patterns in input data and generate new content that shares those patterns but with unique elements.

The property and casualty insurance industry is one area where this technology could have a profound impact.

According to a recent report by Enterprise Apps, a  software that automates business processes, the generative AI market for the insurance industry is expected to reach $5.5 billion by 2032. Companies are increasingly investing in this technology to improve their operations and enhance the customer experience. The report predicts that automation of business processes and improved risk assessment are among the key factors driving this growth.

According to Laila Beane, the chief marketing officer of IntellectAI, an insurance technology company, generative AI can help insurers to manage a range of external influences, including economic, regulatory, and environmental factors. Beane explained that insurers are looking to this technology to help them adapt to a rapidly changing environment, and better meet the needs of their customers and employees.

 “Unlike most widely available AI models and techniques in use for decades, generative AI provides the ability to analyse vast volumes and variety of data, to generate actionable insights for enhancing decision-making processes, personalising customer and broker experiences, providing dynamic loss reserving, performing reputation monitoring, recommending dynamic pricing and coverage options and more,” she said.

Beane said that generative AI-powered chatbots can offer personalised experiences to customers and brokers by answering questions, providing summaries, guiding them through claims, and suggesting products. She further stated that the technology can analyse large amounts of text data to provide valuable insights for risk assessment, claims processing, and personalised coverage, which can help to improve customer retention and boost business growth without increasing workloads for employees.

“By analysing data related to an individual’s behaviour, lifestyle, medical and social history, genetic data points, and more, generative AI can help insurers personalise pricing models that reflect a customer’s actual risk profile more accurately,”she noted.

Beane believes that underwriting is the perfect place to apply generative AI because it’s a data-heavy process that requires unbiased decision-making. She opined that generative AI can help streamline processes, analyze large amounts of data, and recommend the best-fit risks for the business,adding that it also helps reduce bias in decision-making and improve personalized communications, cross-selling, and upselling opportunities.Beane thinks that underwriting is where this technology can have the biggest impact.

Blake Hill, vice president of North America Sales at dacadoo, believes that generative AI can be integrated into global underwriting programmes to reduce human error, streamline the process, and improve accuracy.

Hill said that this technology could be used to summarize lengthy documents, compare information from different providers, and extract data from documents, all of which would reduce the workload for human underwriters. He also suggested that generative AI could be used to write feedback for advisors and customers regarding the underwriting process.

According to Eileen Potter, vice president of insurance marketing at Smart Communications, the companies that take advantage of the benefits of generative AI will gain a competitive edge. Potter believes that this technology can help insurers analyse large amounts of data quickly and extract insights that humans might not see, leading to more informed underwriting decisions and improved risk strategies. He noted further that generative AI can also automate some of the underwriting process, allowing human underwriters to focus on more complex cases. In short, Potter believes that the future of underwriting will be heavily impacted by the power of AI.

However, some potential drawbacks of using generative AI in insurance include a lack of transparency and potential bias in the algorithms. Because the algorithms are complex and the inner workings are difficult to understand, it can be challenging to ensure that they are operating fairly and accurately. There is also a risk that the algorithms could be trained on biased historical data, leading to unfair treatment of certain groups of people. Finally, there is the potential for job loss for humans who currently perform tasks that could be automated by AI.

Beane believes that while generative AI has the potential to transform the insurance industry, some insurers may be reluctant to adopt it due to concerns about data privacy, job loss, and regulatory compliance. However, she stresses that insurers should not be put off by these challenges, and should start with small projects that have clear goals and outcomes. Additionally, Beane emphasizes the importance of change management practices, communication, and staff training in order to ensure the successful implementation of AI solutions. Without these elements, staff resistance can be a major barrier to adoption.

Beane suggests that insurers should use generative AI in a hybrid approach where AI serves as an intelligent assistant rather than replacing human workers. She believes that the most successful AI implementations are not just applied to existing processes, but rather aim to improve processes while still incorporating human involvement. This approach can help ensure that critical human checkpoints are not overlooked, while also eliminating any unnecessary steps in the process.