Generative artificial intelligence (GenAI) is set to transform the Australian insurance sector, driving improvements in efficiency and effectiveness, according to Puneet Bajpai, head of the Xceedance Centre of Excellence for Generative AI.
Bajpai – alongside Prateek Vijayvergia, Xceedance business leader, key accounts in Australia – introduced key concepts of AI, GenAI, and large language models (LLMs) to brokers and underwriting agency staff in an Underwriting Agencies Council webinar.
Vijayvergia emphasised that many tasks in the insurance industry are complex and time-consuming, but AI, GenAI, and LLMs can quickly process and analyse vast amounts of data, replicating human-like functions.
Bajpai explained the difference between traditional AI and GenAI. Traditional AI is a focused tool designed for specific use cases. By contrast, GenAI is capable of performing a variety of tasks and generating content by utilising data from trillions of words available online, based on prompts.
“When we train a large AI system on a lot of data, we create an LLM. The introduction of a new architecture called transformers enabled global entities, like Google, Meta, Microsoft, and others, to use transformers to train models on gigantic text datasets,” he said. “For an LLM to know what is being asked of it, the user must make an entry, called a prompt. The better the human input, the better the LLM output. LLMs are not yet doing the thinking for users.”
Bajpai also noted the importance of prompt engineering, which involves optimising textual input to finetune LLM responses, allowing tasks to be completed in seconds rather than weeks.
Focusing on underwriting agencies, both experts said GenAI can improve customer experience through data-driven product design, virtual assistants, and personalised communications with both customers and brokers.
The technology can also enhance operational efficiency via product configuration, submission ingestion, and quote generation and provide business insights for customer targeting and risk assessment and rating.
Vijayvergia expressed his enthusiasm for the potential of GenAI and LLMs in the insurance industry.
“The ability to remove cumbersome tasks enables agencies and brokers to spend more time with clients to better understand their risks,” he said. “GenAI, being agnostic, can remove the pain points of working across different platforms and use cases.”
He also noted that while the broader insurance industry has been slow to adopt new technology, underwriting agencies have an opportunity to change that.
“Most agencies are relatively new, without legacy tech and with simpler decision-making processes. Those that choose to embrace AI could soon be way ahead of their competitors,” he said.
