AI in insurance
The current times where natural disasters are increasing in frequency and severity, traditional insurance models struggle to keep up. Insurers today face unprecedented challenges, particularly in accurately predicting losses associated with catastrophic events such as wildfires. The good part is that technologies like artificial intelligence (AI) and geographic information systems (GIS), combined with data, are redefining catastrophe modeling, resulting in more precise risk assessments and pricing strategies.
A Senior Data Scientist, Jwalin Thaker, is assisting this transformation by using these advanced tools to revolutionize insurance pricing and risk management. He shared how his journey into modernizing insurance practices began during his graduate studies in Applied Artificial Intelligence, where he explored leveraging AI to refine and enhance traditional actuarial pricing strategies and optimize policy admin systems. His early successes included developing sub-second latency cloud-based APIs that made quoting multiple risks (or portfolio) at scale very efficient, paving the way for his rapid professional advancement.
Thaker's work stands out in building models for areas prone to severe natural disasters. By using advanced machine learning methods, like gradient boosted trees and generalized additive models (GAMs), and combining them with geospatial analysis and GIS data, he created smarter underwriting and rating models. His hurricane models for the U.S. East Coast greatly improved pricing accuracy and gave the insurer a stronger market position, leading to a five-to tenfold growth in written premiums since 2022.
Building on these achievements, he turned his attention to California's escalating wildfire risk. He harnessed rich environmental datasets combined with detailed GIS mapping to construct a strong wildfire frequency prediction model. The success of this approach was dramatically demonstrated during the January 2025 Los Angeles fires, where policies underwritten with the professional's models saw zero filed claims saving the firm over $300K in claims processing fees and over $30M in potential probable losses. These results showed how effective his improved underwriting grids and accurate pricing models were. He used machine learning, historical data, and forecasts of wildfire risks to validate them. His focus on transparency and easy-to-understand methods gained strong support from both underwriters and regulators.
Beyond catastrophe modeling, Thaker also pioneered advancements in claims automation. He developed an AI-driven claims detection platform capable of automating the triaging process, significantly reducing manual labor and projected to save over $1 million annually.
Further expanding the role of AI in insurance, he recently created a Generative AI chatbot utilizing a sophisticated Retrieval-Augmented Generation (RAG) architecture. This innovative chatbot effectively addresses frequent inquiries regarding product features, pricing, inspection guidelines, and claims processes. Its deployment is anticipated to double customer service operational efficiency and drastically reduce response times.
Drawing from the advancements shared in this article, it's clear how the insurance industry is entering a new era where technology plays a crucial role in managing risk. AI, GIS data, and advanced ML models are making loss prediction more accurate and helping insurers respond faster to catastrophic events. As natural disasters grow in scale and unpredictability, these technologies are not just improving pricing strategies but also enabling stronger financial resilience and better protection for communities.
