Beyond expert users: agents should help users construct preferences, not just elicit them
Summary
This paper argues that agents should help users construct preferences rather than assuming well-formed ones, proposing the CoPref model and CoShop benchmark. Evaluations show even frontier models achieve only 56% accuracy due to poor preference expansion.
View Cached Full Text
Cached at: 07/01/26, 05:36 AM
# Beyond expert users: agents should help users construct preferences, not just elicit them Source: [https://arxiv.org/abs/2606.30863](https://arxiv.org/abs/2606.30863) [View PDF](https://arxiv.org/pdf/2606.30863) > Abstract:Agents typically assume an expert user \-\- one with well\-formed preferences about what they want \-\- and default to clarifying questions whenever the task is underspecified\. We argue this assumption is unrealistic\. Users often lack the domain knowledge to have completely specified preferences; if asked about their preference on some feature, the user may be unable to answer without the agent helping the user to learn some domain knowledge needed to form a preference for that feature, e\.g\., via examples or explanations\. To formalize these principles, we draw on the Search\-Experience\-Credence framework from Information Economics to introduce CoPref, a model of how users construct preferences based on agent dialog actions\. We then study these ideas concretely in agentic recommender systems, proposing CoShop, an interactive benchmark\. In CoShop, an agent converses with and makes recommendations for a CoPref user\. The agent's performance depends on whether it can help the user gain the knowledge needed to specify the task well\. Evaluating five frontier models, we find that no agent exceeds 56% accuracy on CoShop despite five turns of interaction\. Failures stem not from agents' ability to find items, but from how little the interaction expands what users know about what they want\. ## Submission history From: Irena Saracay \[[view email](https://arxiv.org/show-email/a7f20c4b/2606.30863)\] **\[v1\]**Mon, 29 Jun 2026 19:50:16 UTC \(2,270 KB\)
Similar Articles
iPrefBench: A User-Centric Benchmark for Personalized Mobile Use Agents
iPrefBench is a user-centric benchmark evaluating whether personal mobile-use agents can infer missing preferences from activity history or ask for clarification in an executable iOS environment with a simulated user. Evaluation of 7 frontier agents shows a large gap: high action completion (up to 91.6%) but low full task success (up to 47.7%), with agents especially struggling to request necessary ask-only information.
Knowledge Agents: Beat Frontier Models with Better Structure (18 minute read)
The article presents 'knowledge agents', a methodology that injects relevant knowledge into AI agents via a hybrid retrieval system, allowing smaller models to outperform large frontier models across specialized domains like financial markets, policy, and healthcare.
When More Data Is Not Enough: The Context-Sufficiency Frontier in Generative AI Personalization
This arXiv preprint develops a theory of context sufficiency for generative AI personalization, proposing a Context-Sufficiency Frontier that distinguishes four states (insufficiency, sufficiency, saturation, interference) and showing via a full-factorial experiment with a generative recommender at a home-furnishing retailer that relevant context improves appropriateness while irrelevant context harms it and destabilizes retrieval.
Prompt Optimization for User Simulation in Conversational Recommender Systems: A Multi-Objective Framework
This paper proposes a framework to automatically optimize prompts for LLM-based user simulators in conversational recommender systems, addressing issues like positive bias and limited behavioral diversity.
Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It
This paper studies how LLM agents exhibit source preference in end-to-end search, favoring items from certain sites even when they meet fewer requirements, and demonstrates that supplying missing information or counter-prompting can reduce this bias across 12 agent models.