Abstract

As the volume of peer-reviewed research surges, scholars increasingly rely onsocial platforms for discovery, while authors invest considerable effort inpromoting their work to ensure visibility and citations. To streamline thisprocess and reduce the reliance on human effort, we introduce AutomaticPromotion (AutoPR), a novel task that transforms research papers into accurate,engaging, and timely public content. To enable rigorous evaluation, we releasePRBench, a multimodal benchmark that links 512 peer-reviewed articles tohigh-quality promotional posts, assessing systems along three axes: Fidelity(accuracy and tone), Engagement (audience targeting and appeal), and Alignment(timing and channel optimization). We also introduce PRAgent, a multi-agentframework that automates AutoPR in three stages: content extraction withmultimodal preparation, collaborative synthesis for polished outputs, andplatform-specific adaptation to optimize norms, tone, and tagging for maximumreach. When compared to direct LLM pipelines on PRBench, PRAgent demonstratessubstantial improvements, including a 604% increase in total watch time, a 438%rise in likes, and at least a 2.9x boost in overall engagement. Ablationstudies show that platform modeling and targeted promotion contribute the mostto these gains. Our results position AutoPR as a tractable, measurable researchproblem and provide a roadmap for scalable, impactful automated scholarlycommunication.