🤖 AI Summary
This study addresses the limitations of traditional actuarial pricing, which relies heavily on manual feature engineering and struggles to leverage unstructured textual data. It proposes a novel approach that integrates pretrained large language models (LLMs) into the insurance pricing pipeline by extracting deterministic embeddings from policyholders’ natural language descriptions via controlled prompting. These embeddings are then incorporated as input features in a generalized linear model for Poisson regression of claim frequency. The method significantly reduces dependence on handcrafted features while preserving model governability. Empirical evaluation on French motor third-party liability insurance data demonstrates that the embedding-driven model substantially outperforms conventional approaches in low-data regimes. In high-data settings, performance is influenced by model architecture and embedding dimensionality, with further gains achieved through domain-specific fine-tuning.
📝 Abstract
Classical actuarial pricing models, such as the generalized linear model, are valued for transparency and ease of governance, but they use interactions among risk factors only when these are supplied through explicit feature engineering. We study whether embeddings from a pre-trained large language model, computed from a natural-language description of each policyholder, can replace hand-crafted features as inputs to a standard actuarial pricing model, taking Poisson claim-frequency regression as the main example. The language model is used only to construct deterministic embedding covariates; pricing is performed by a standard generalized linear model. Using French motor third-party liability data, the embedding-based model outperforms the generalized linear model, especially when data are scarce, whereas at larger sample sizes the comparison is model- and dimension-dependent. Insurance-specific fine-tuning further improves the embeddings, and a prompt-sensitivity diagnostic shows that the pipeline reacts to any appended out-of-template field, making controlled prompts a governance requirement.