Researchers have developed a Large Behavioral Model (LBM) that acts as a promptable digital twin for retail customers. This model learns customer decision-making directly from large-scale transaction data using a Person-Environment formulation. The LBM incorporates customer state through historical purchase profiles and product context via retrieval-augmented generation. It is trained using continued pre-training, supervised fine-tuning, and reinforcement learning, demonstrating superior performance over general-purpose language models on various retail tasks and showing strong transfer capabilities across different retailers and decision domains. AI
IMPACT This model could significantly enhance personalization and decision support in retail by providing a more accurate and explainable digital twin of customer behavior.
RANK_REASON This is a research paper detailing a new model and methodology. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
