KnowU-Bench: Towards Interactive, Proactive, and Personalized Mobile Agent Evaluation

TL;DR

KnowU-Bench is a comprehensive, interactive Android-based benchmark designed to evaluate personalized mobile agents' ability to infer preferences, make proactive decisions, and handle real-world user interactions, addressing gaps in existing static benchmarks.

Contribution

It introduces a novel, interactive benchmark with a user simulator, covering preference elicitation, consent negotiation, and proactive assistance, enabling realistic evaluation of personalized mobile agents.

Findings

Agents perform below 50% accuracy on vague instructions requiring preference inference.

Preference acquisition and intervention calibration are the main bottlenecks.

Current models struggle with trustworthiness and proactive decision-making in realistic scenarios.

Abstract

Personalized mobile agents that infer user preferences and calibrate proactive assistance hold great promise as everyday digital assistants, yet existing benchmarks fail to capture what this requires. Prior work evaluates preference recovery from static histories or intent prediction from fixed contexts. Neither tests whether an agent can elicit missing preferences through interaction, nor whether it can decide when to intervene, seek consent, or remain silent in a live GUI environment. We introduce KnowU-Bench, an online benchmark for personalized mobile agents built on a reproducible Android emulation environment, covering 42 general GUI tasks, 86 personalized tasks, and 64 proactive tasks. Unlike prior work that treats user preferences as static context, KnowU-Bench hides the user profile from the agent and exposes only behavioral logs, forcing genuine preference inference rather than context lookup. To support multi-turn preference elicitation, it instantiates an LLM-driven user simulator grounded in structured profiles, enabling realistic clarification dialogues and proactive consent handling. Beyond personalization, KnowU-Bench provides comprehensive evaluation of the complete proactive decision chain, including grounded GUI execution, consent negotiation, and post-rejection restraint, evaluated through a hybrid protocol combining rule-based verification with LLM-as-a-Judge scoring. Our experiments reveal a striking degradation: agents that excel at explicit task execution fall below 50% under vague instructions requiring user preference inference or intervention calibration, even for frontier models like Claude Sonnet 4.6. The core bottlenecks are not GUI navigation but preference acquisition and intervention calibration, exposing a fundamental gap between competent interface operation and trustworthy personal assistance.