NExT-Search: Rebuilding User Feedback Ecosystem for Generative AI Search

TL;DR

NExT-Search proposes restoring a fine-grained feedback ecosystem for generative AI search by introducing User Debug Mode and Shadow User Mode to inject stage-level supervision into query decomposition, retrieval, and answer generation. It couples online adaptation for real-time refinements with offline updates for periodic model improvements, and introduces a Feedback Store to incentivize user participation. The framework aims to combine the convenience of end-to-end generation with continuous, interpretable feedback signals to sustain self-improvement in AI search. This paradigm addresses a core gap in current generative search systems by enabling iterative optimization across pipeline components while preserving user experience.

Abstract

Generative AI search is reshaping information retrieval by offering end-to-end answers to complex queries, reducing users' reliance on manually browsing and summarizing multiple web pages. However, while this paradigm enhances convenience, it disrupts the feedback-driven improvement loop that has historically powered the evolution of traditional Web search. Web search can continuously improve their ranking models by collecting large-scale, fine-grained user feedback (e.g., clicks, dwell time) at the document level. In contrast, generative AI search operates through a much longer search pipeline, spanning query decomposition, document retrieval, and answer generation, yet typically receives only coarse-grained feedback on the final answer. This introduces a feedback loop disconnect, where user feedback for the final output cannot be effectively mapped back to specific system components, making it difficult to improve each intermediate stage and sustain the feedback loop. In this paper, we envision NExT-Search, a next-generation paradigm designed to reintroduce fine-grained, process-level feedback into generative AI search. NExT-Search integrates two complementary modes: User Debug Mode, which allows engaged users to intervene at key stages; and Shadow User Mode, where a personalized user agent simulates user preferences and provides AI-assisted feedback for less interactive users. Furthermore, we envision how these feedback signals can be leveraged through online adaptation, which refines current search outputs in real-time, and offline update, which aggregates interaction logs to periodically fine-tune query decomposition, retrieval, and generation models. By restoring human control over key stages of the generative AI search pipeline, we believe NExT-Search offers a promising direction for building feedback-rich AI search systems that can evolve continuously alongside human feedback.

Figures (3)

• Figure 1: Comparison of the paradigm of traditional web search engines and generative AI search engines. (a) Web search retrieves and ranks results, presenting them as a Search Engine Results Page (SERP), where user feedback on the document level can be directly leveraged to update the ranking model. (b) Generative AI search automates multiple steps to generate direct answers, but its extended pipeline complicates the effective use of user feedback for refining individual components.
• Figure 2: Illustration of our proposed NExT-Search paradigm. NExT-Search introduces a dual feedback mechanism to enhance generative AI search: In User Debug Mode, users can intervene at different stages—query decomposition, retrieval, and answer generation—to refine search results with granular feedback. In Shadow User Mode, a personalized agent simulates user behavior to assist in providing feedback with minimal user interaction, reducing user engagement costs.
• Figure 3: Two mechanisms for leveraging feedback.