[目的/意义] 纸质文献采访仍是当前图书馆最重要的基础工作之一，基于读者需求的不确定性和图书价值评估，采访工作的当务之急是在保证质量的前提下提高采访效率。需求与决策驱动的图书智能采访系统旨在以高效智能的思维方式，融入科学的资源采访机制，以提高采访工作效率和馆藏资源质量为目标，实现图书馆采访工作的智能化。[方法/过程] 在调研智能化采访的研究与实践现状的基础上，根据学校学科建设需求、出版社模型、读者行为分析、读者推荐以及图书价格等多种维度，设置业务规则及决策权重，让系统智能化筛选纸质文献。以重庆大学图书馆智能采访系统为例，分析系统内置的图书评价、出版社与学科质量、作者与学科质量3个模型，优化采访工作业务流程，并实现新书征订、图书增补、读者荐购等不同业务场景分别制定相应策略，优化采访工作流程。[结果/结论] 图书智能采访系统通过读者需求和重点学科建设决策驱动，在人工智能技术的支撑下，能有效提高纸质文献资源建设工作的效率和质量，将成为当前采访模式变革的主要方向。

[Purpose/significance] Paper literature acquisitioning is still one of the most important basic tasks of the current library. Faced with the uncertain needs of readers and the evaluation of book value, the urgent task of interviewing is to improve the efficiency of interviews under the premise of ensuring quality. The decision-driven book intelligent interview system aims to integrate the scientific resource interview mechanism with an efficient and intelligent way of thinking to improve the efficiency of interview work and the quality of collection resources, and to realize the intelligentization of library interview work. [Method/process] Based on the research and practice status of the intelligent interviews, and according to the school discipline construction needs, the publisher model, the reader behavior analysis, the reader recommendation and the book price, the business rules and the weight of decision-making were set, which allowing the system to intelligently screen paper documents. This paper used the intelligent interview system of Chongqing University Library as an example to analyze 3 models of book evaluation, publishing house and subject quality, author and subject quality, and optimized interview work. Process, and implement different strategies for new book subscription, book addition, reader recommendation and other business scenarios to optimize the interview workflow. [Result/conclusion] This paper argues that the book intelligent interview system is driven by readers' requirement and "key disciplines" construction decision-making. Under the support of artificial intelligence technology, the construction of paper literature resources will effectively improve efficiency and quality, and will definitely become the current interview. The main direction of model change.