PAARS: Persona Aligned Agentic Retail Shoppers

2025-03-31

AI Summary

Large language model (LLM)-driven shopping agents exhibit systematic deviations from real user behavior in brand preference, rating bias, and demographic representativeness. Method: We propose the first synthetic agent modeling framework explicitly designed for population-level distribution alignment. It (1) automatically discovers fine-grained, anonymized user personas from historical behavioral data; (2) equips agents with retail-specific toolkits to enable controllable, task-aware dialogue generation; and (3) introduces a suite of group-level consistency metrics based on Wasserstein distance and KL divergence. Contribution/Results: This work pioneers alignment at the population-distribution level—beyond individual behavioral fidelity. It unifies data-driven persona extraction and agent instantiation into a single pipeline. End-to-end evaluation demonstrates that our agent population accurately reproduces real-world A/B test trends, achieving strong validity and generalization capability in commercial decision simulation.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search

Abstract

In e-commerce, behavioral data is collected for decision making which can be costly and slow. Simulation with LLM powered agents is emerging as a promising alternative for representing human population behavior. However, LLMs are known to exhibit certain biases, such as brand bias, review rating bias and limited representation of certain groups in the population, hence they need to be carefully benchmarked and aligned to user behavior. Ultimately, our goal is to synthesise an agent population and verify that it collectively approximates a real sample of humans. To this end, we propose a framework that: (i) creates synthetic shopping agents by automatically mining personas from anonymised historical shopping data, (ii) equips agents with retail-specific tools to synthesise shopping sessions and (iii) introduces a novel alignment suite measuring distributional differences between humans and shopping agents at the group (i.e. population) level rather than the traditional"individual"level. Experimental results demonstrate that using personas improves performance on the alignment suite, though a gap remains to human behaviour. We showcase an initial application of our framework for automated agentic A/B testing and compare the findings to human results. Finally, we discuss applications, limitations and challenges setting the stage for impactful future work.

Problem

• Simulating human shopping behavior using LLM agents with reduced biases
• Aligning synthetic agent populations to real human behavior distributions
• Automating A/B testing via persona-based shopping agent simulations

Innovation

• Automatically mines personas from historical shopping data
• Equips agents with retail-specific shopping tools
• Measures alignment at population level, not individual