Unlocking the Potential: Assessing Large Language Models for the Insurance Industry

As the insurance industry embraces digital transformation, cutting-edge technologies are reshaping the way we approach challenges and opportunities. Among these, Large Language Models (LLMs) have emerged as a powerful and promising innovation with vast potential to revolutionize various aspects of the insurance business.

Join us in this thought-provoking talk as we delve into the exciting world of Large Language Models and explore their capabilities within the insurance domain.

Through a comprehensive comparative analysis, we focus on two leading open-sourced LLMs, Falcon and Vicuna, to explore how these advanced models can transform various insurance-related tasks.

Central to our assessment is a set of carefully crafted insurance domain questions, covering a wide range of question types and aspects within the insurance field. This curated list allows for a comprehensive evaluation of how well the large language models understand and perform in insurance-specific contexts. By adopting a multi-dimensional evaluation criterion, we gain insights into the strengths and potential limitations of Falcon and Vicuna models, assessing their handling of insurance-related inquiries from various angles.

Throughout the talk, we will present domain examples and evaluation scenarios, showcasing how each model shines in different contexts and settings within the insurance domain. Additionally, we will outline a few potential use cases and recommend a suitable LLM based on their respective strengths. Get ready to explore the uncharted territories of Large Language Models in insurance and unlock their transformative potential.

Whether you are an industry expert, data enthusiast, or technology advocate, this talk will ignite your curiosity and spark engaging discussions on the future of insurance. Together, we’ll navigate through the untapped opportunities that await in this exciting era of Generative AI.


About the speaker

Ke Ji

Senior Lead Data Scientist at American Family Insurance

Meet Ke, Senior Lead Data Scientist at American Family Insurance, driven by a passion for crafting transformative data-driven solutions that fuel technical innovation and business growth.

She plays a vital role in shaping the technical strategies and cultivating innovation within the data science and engineering team, guiding the design and execution of AI solutions that bridge short-term business needs with long-term vision. With over 7 years of data science mastery,

Ke’s journey is marked by her innate ability to unravel data-driven complexities, formulate robust data science strategies, and build scalable end-to-end ML applications tailored for production in insurance business.

Holding a Master’s degree in Statistics from UIUC and a B.S. in Statistics from Renmin University in China, her proficiency spans machine learning model development, predictive analytics, advanced statistics, ingenious problem-solving for intricate business challenges, and particularly shines in the realm of Natural Language Processing.




Online Event: October 3-5, 2023




Presented by