Modern e-commerce search is evolving from simple keyword matching to resolving complex user intents. While large language models (LLMs) offer powerful reasoning capabilities, existing LLM-based search paradigms suffer from a fundamental blindness-latency dilemma: query rewriting methods are agnostic to retrieval tool capabilities and real-time inventory states, resulting in invalid plans; conversely, deep search agent approaches initially plan without environment awareness, then rely on iterative tool calls and reflection to perceive and correct failures, leading to seconds of latency, incompatible with the sub-second budget for the planning module in industrial e-commerce search. To resolve this conflict, we propose Environment-Aware Search Planning (EASP), a novel paradigm that reformulates search planning as a dynamic reasoning process grounded in environmental reality. EASP introduces a Probe-then-Plan mechanism: a lightweight Retrieval Probe first exposes the retrieval snapshot, enabling the Planner to diagnose execution gaps and generate grounded search plans. Our methodology unfolds in three stages: (1) Offline Data Synthesis: The Teacher Agent synthesizes diverse, execution-validated plans by diagnosing the retrieval environment exposed by the Retrieval Probe. (2) Planner Training and Alignment: The Planner is initialized via supervised fine-tuning (SFT) on the offline dataset to internalize the Teacher's diagnostic capabilities, followed by alignment with business outcomes (conversion rate) through reinforcement learning. (3) Adaptive Online Serving: A complexity-aware routing mechanism selectively activates the planning pipeline only for complex queries, ensuring optimal resource allocation. Extensive offline evaluations and online A/B testing on JD.com demonstrate that EASP significantly improves relevant recall and achieves substantial lifts in UCVR and GMV. EASP has been successfully deployed in JD.com's AI-Search system.

Faceted search acts as a critical bridge for navigating massive e-commerce catalogs, yet traditional systems rely on static rule-based extraction or statistical ranking, struggling with emerging vocabulary, semantic gaps, and a disconnect between facet selection and underlying retrieval. In this paper, we introduce GenFacet, an industrial-grade, end-to-end generative framework deployed at JD.com. GenFacet reframes faceted search as two coupled generative tasks within a unified Large Language Model: Context-Aware Facet Generation, which dynamically synthesizes trend-responsive navigation options, and Intent-Driven Query Rewriting, which translates user interactions into precise search queries to close the retrieval loop. To bridge the gap between generative capabilities and search utility, we propose a novel multi-task training pipeline combining teacher-student distillation with GRPO. This aligns the model with complex user preferences by directly optimizing for downstream search satisfaction. Validated on China's largest self-operated e-commerce platform via rigorous offline evaluations and online A/B tests, GenFacet demonstrated substantial improvements. Specifically, online results reveal a relative increase of 42.0% in facet Click-Through Rate (CTR) and 2.0% in User Conversion Rate (UCVR). These outcomes provide strong evidence of the benefits of generative methods for improving query understanding and user engagement in large-scale information retrieval systems.