We present a modular two-agent simulation framework for evaluating conversational shopping assistant architectures. An independent buyer agent, configured with personas, missions, and patience levels, is paired with an interchangeable responder that integrates with a real e-commerce search API. Holding the buyer constant across experiments enables controlled comparison of responder designs on identical scenarios. Using 2011 conversations across 14 persona buckets, we establish four empirical findings. First, rolling-window memory outperforms intent-extraction memory on all quality metrics while being 35% faster per query. Second, illustrating rapid evidence-driven iteration, a systematic failure analysis of a responder version enables targeted fixes that reduce failure and near-failure rates by 62% across the full dataset. Third, swapping the responder LLM backbone from Gemini~2.5 to Llama~3.3~70B costs 0.16--0.45 points despite identical architecture. Finally, we document systematic philosophical disagreement between frontier LLM judges: Gemini rewards process correctness while Claude demands concrete outcomes, despite using the same evaluation prompt.

翻译：我们提出一种模块化双智能体仿真框架，用于评估对话式购物助手的架构。该框架将配置有人设、任务及耐心等级的独立买家智能体，与一个可替换的响应器配对，该响应器集成于真实的电商搜索API。通过在实验中固定买家变量，我们能够在相同场景下对响应器设计方案进行控制性比较。基于涵盖14个人设类别的2011段对话，我们得出四项实证发现：第一，滑动窗口记忆在所有质量指标上均优于意图提取记忆，且每次查询速度提升35%；第二，通过对某个响应器版本的系统性故障分析，我们实现了基于证据的快速迭代——针对性的修复措施使整个数据集的失败与濒临失败率降低62%；第三，将响应器的大语言模型骨干从Gemini~2.5替换为Llama~3.3~70B时，尽管架构完全相同，性能仍下降0.16-0.45个点；最后，我们发现前沿大语言模型裁判间存在系统性哲学分歧：即使使用相同的评估提示，Gemini倾向于评价过程正确性，而Claude则强调具体结果。