Beyond automation: moving to generative AI – Xceedance

Artificial intelligence (AI) is the latest innovation in the digital transformation of the insurance industry – and the most talked about.

There’s no doubt that AI speeds up data processing and frees up insurance professionals by automating repetitive tasks. But the industry is now poised to move beyond this basic level of automation and machine learning to discover the full potential of this technology.

Many companies have already made substantial investments in AI to carry out insurance operations such as underwriting, claims processing, customer service and fraud detection. But to see truly game-changing results, insurers must now focus on four areas: in-depth testing, workflow transformation, customization and integration.

Building trust in AI

There is still a great deal of unease among insurers about the challenges thrown up by advancements in AI, including staffing, regulatory compliance, security, and effective integration. Insurers can build trust in AI by using transparent systems that are rigorously tested to avoid inherent bias, safeguard data, and protect privacy, and by ensuring that AI decisions are validated by human operators.

Insurers are currently feeling their way with AI, testing out how an operation carried out by AI compares to traditional processes. As growing numbers of companies adopt AI, the more trust and comfort will be built within the insurance industry.

Transforming workflows

Generative AI transforms workflows by assimilating data from diverse sources and extracting insights into evolving risks such as cyber and climate change. The use of AI tools - such as ChatGPT integrated with Azure OpenAI Service, for example - enables insurers to make quicker, more accurate decisions by rapidly processing and analysing vast amounts of data, revolutionizing risk assessment and underwriting processes.

Claims managers’ and underwriters’ workflows can be made more efficient by AI, which uses large language models and prompt engineering to precisely extract and interpret data from lengthy documents such as claims adjudicator reports and property surveys, reducing the risk of human error and of missing important information.

While AI removes the need for a team to initially collect and collate the data, there still a need for human oversight to provide quality assurance of the distilled information.

A tailored approach

By analysing extensive data sources, GenAI enables insurers to tailor policies precisely to individual circumstances and needs while allowing insurers to better manage risk and price more accurately.

AI enables allows to use data from a wider range of sources than usual, including social media, to tailor coverage to individual risks. For example, if someone is looking for car insurance, AI will mine publicly available information that points to their driving behaviour or any associations that motivate them to drive more recklessly such as an interest in racing cars. AI is particularly helpful in health insurance, where the technology can evaluate personal health data to offer highly personalized policies.

Integration is key

GenAI requires a great deal of data, and the challenge is how to integrate various sources into platforms to enable the software to inform decision-making.

Creating a sandbox environment for GenAI experimentation allows users to explore AI capabilities, identify valuable use cases, and adapt the technology to their needs. In an industry where human expertise is vital for customer engagement, the sandbox approach paves the way for successful integration of GenAI.

Xceedance has set up a Centre of Excellence for Generative AI with a sandbox environment for experimenting with the technology. This approach exposes users to GenAI in a controlled and safe way and allows them to identify high-return and fail fast use cases, fostering technological exploration.

Building a multi-disciplinary team of business experts, IT specialists, and data scientists is essential for the successful integration of AI into organizational processes. The team’s focus should be on adapting AI solutions to meet the company’s needs, starting with low barrier use cases, and progressively advancing to more complex applications. By doing so, insurers can develop the necessary expertise and capabilities to deploy AI effectively and safely

AI’s potential to revolutionise the insurance industry cannot be overstated, but realizing this potential requires a commitment to building trust, understanding the technology, and developing a strategy to implement AI solutions in certain areas while proceeding with caution in others.

By Amit Tiwari (pictured), co-founder and president at Xceedance.