Search components in e-commerce apps, often complex AI-based retrieval systems, are susceptible to bugs, including ones leading to missed recalls. Missed recalls are instances where an entry should appear in the search results based on the algorithmic and business logic, but doesn’t. They cause dissatisfaction among shop owners and can impact the app’s profitability. However, testing for missed recalls is challenging due to difficulties in generating user-aligned test cases and the absence of oracles. In this paper, we introduce mrDetector, the first automatic testing approach specifically for missed recalls. To tackle the test case generation challenge, we first study how users construct queries during searching. We then use these findings to create a CoT prompt containing multiple examples and guide the LLM generation process. To address the lack of oracles, we learn from users who create multiple queries for one shop and compare search results, and provide a test oracle through a metamorphic relation. Extensive experiments using open access data demonstrate that mrDetector outperforms all baselines with the lowest false positive ratio. Experiments with real industrial data show that mrDetector discovers over one hundred missed recalls with only 17 false positives. Corresponding engineers accept all representative missed recalls mrDetector finds.