Grocery to General Merchandise: A Cross-Pollination Recommender using LLMs and Real-Time Cart Context

2025-09-02

AI Summary

This paper addresses the underexplored yet critical problem of cross-category general-item recommendation in grocery shopping scenarios. We propose XP, a cross-category recommendation framework that jointly models co-purchase behavior, implicit semantic associations (e.g., “milk → milk frother”) extracted by large language models (LLMs), and real-time sequential cart context within a Transformer-based dynamic ranking architecture. Our key contributions are twofold: (1) the first application of LLMs to discover fine-grained, interpretable cross-category item relationships; and (2) the realization of cart-level, real-time contextual awareness in recommendation. Offline evaluations demonstrate that LLM-enhanced modeling improves add-to-cart rate by 36%; online A/B testing shows XP achieves a 27% gain in NDCG@4 over strong baselines, significantly advancing both accuracy and practical utility of cross-category recommendations.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Modern e-commerce platforms strive to enhance customer experience by providing timely and contextually relevant recommendations. However, recommending general merchandise to customers focused on grocery shopping -- such as pairing milk with a milk frother -- remains a critical yet under-explored challenge. This paper introduces a cross-pollination (XP) framework, a novel approach that bridges grocery and general merchandise cross-category recommendations by leveraging multi-source product associations and real-time cart context. Our solution employs a two-stage framework: (1) A candidate generation mechanism that uses co-purchase market basket analysis and LLM-based approach to identify novel item-item associations; and (2) a transformer-based ranker that leverages the real-time sequential cart context and optimizes for engagement signals such as add-to-carts. Offline analysis and online A/B tests show an increase of 36% add-to-cart rate with LLM-based retrieval, and 27% NDCG@4 lift using cart context-based ranker. Our work contributes practical techniques for cross-category recommendations and broader insights for e-commerce systems.

Problem

• Recommending general merchandise during grocery shopping
• Leveraging multi-source product associations and real-time cart context
• Bridging cross-category recommendations between grocery and merchandise

Innovation

• LLM-based item association retrieval
• Transformer ranker with cart context
• Two-stage cross-category recommendation framework