承诺视角下企业对人工智能推荐服务的策略性使用

doi:10.16381/j.cnki.issn1003-207x.2023.1888

摘要/Abstract

摘要：

目前，基于人工智能（AI）的个性化推荐服务已经被企业广泛应用于市场营销的各个方面。企业利用AI对消费者的购买行为进行分析，通过用户画像了解消费者偏好，在未来的销售季度能够智能推荐更加符合消费者偏好的商品。一方面，AI可以帮助提升推荐商品和消费者偏好之间的匹配度，以刺激消费进而扩大市场，形成对销售的促进效应；另一方面，也可能导致策略性消费者形成未来会被推荐更合适商品的预期，进而助长观望情绪，由此带来延迟购买效应，这使得企业在利用AI进行个性化推荐时，变得更加谨慎。本文从企业采取的承诺策略这一视角出发，研究企业应该如何策略性地使用基于AI的个性化推荐服务。通过构建两期的动态博弈模型，本文考察两种承诺策略：（1）价格承诺策略，即企业提前告知两期商品的销售价格，对应现实中商家的价格保护策略等；（2）匹配承诺策略，即企业通过宣传自己的AI推荐，让消费者对未来被推荐商品的匹配精准度提升有正确的预期。承诺策略影响消费者对产品不同销售期的价值判断，进而调节购买行为，为AI推荐的实施形成策略性交互。研究结果表明，引入基于AI的个性化推荐服务，并不一定总是对企业有利，只有当原本的延迟购买效应较弱，或者AI带来的销量提升效应足够强时，AI推荐才能使企业受益。匹配承诺策略看似会加重消费者的观望情绪，促进延迟购买，实际上，反而会减弱该效应，而且，在这一策略下，企业只需提供较低的匹配提升预期。另外，企业对AI推荐的使用程度和其为匹配精度提升所付出的成本之间不一定是正相关的，即可能成本虽然高，企业反而应该大力推广AI推荐，提升商品匹配精度，这取决于在销量提升和延迟购买两个效应上的权衡。当提高匹配精度的成本相对较低时，匹配承诺策略相比于传统的价格承诺更具有优势。本文还比较了不同策略下的消费者剩余和总社会福利。

关键词: 人工智能推荐, 承诺, 匹配精度, 延迟购买, 博弈

Abstract:

AI-based personalized recommendation services are now widely used by companies in various marketing areas. Companies analyze consumers’ purchasing behavior using AI and understand their preferences through user profiling. It allows them to recommend better-match products to consumers in future sales periods. AI can improve matching accuracy between recommended products and consumers， stimulate consumption thus expanding the market and promoting sales. However， it also may lead strategic consumers to expect better product recommendations in the future， wait and delay their purchases， thus resulting profit loss. As a result， companies need to be more cautious using AI-based recommendations. How companies should strategically use AI-based personalized recommendation services is explored through commitment strategies. A two-period dynamic game model is used to explore two commitment strategies：（1）Price commitment strategy， where the company announces the product prices of two periods in advance， similar to real-world price protection strategies；（2）Enhanced-match commitment strategy， where companies promote their AI recommendations to ensure that consumers have accurate expectations about the enhanced-match precision of future recommended products. Commitment strategies influence consumers' value assessments of products across two sales periods， thereby regulating purchasing behavior and creating strategic interactions for AI-based recommendations. The results indicate that AI-based personalized recommendation service is not necessarily beneficial. Only when the originally existence of purchase delay is not severe or the sales-promotion effect is strong enough， can the adoption of AI benefit the firms. Interestingly， although Enhanced-match strategy may seem to encourage delayed purchases， it actually reduces this effect and requires only a modest improvement in matching accuracy. Additionally， the relationship between the AI use extent and the costs of enhanced-match accuracy is not necessarily positive. Companies may need to heavily promote AI-based recommendations to enhance product matching accuracy， even with high costs， depending on the trade-off between sales promotion and delayed purchase effects. When the cost of enhanced-match accuracy is relatively low， the Enhanced-match strategy plays better than traditional price commitment strategies. The consumer surplus and total social welfare are also compared under different strategies.

Key words: AI recommendation, commitment, matching accuracy, delayed purchase, game

中图分类号:

F272

刘慧慧,张浩哲,刘霖. 承诺视角下企业对人工智能推荐服务的策略性使用[J]. 中国管理科学, 2026, 34(3): 345-356.

Huihui Liu,Haozhe Zhang,Lin Liu. The Strategic Use of AI Recommendation Service from the Perspective of Commitment[J]. Chinese Journal of Management Science, 2026, 34(3): 345-356.

图/表 6

图1

图2

图3

图4

图5

图6

参考文献 41

[1]	Panniello U， Hill S， Gorgoglione M. The impact of profit incentives on the relevance of online recommendations［J］. Electronic Commerce Research and Applications， 2016， 20： 87-104.
[2]	Tam K Y， Ho S Y. Web personalization as a persuasion strategy： An elaboration likelihood model perspective［J］.Information Systems Research， 2005，16（3）： 271-291.
[3]	Eslami S P， Hassanein K. Understanding data analytics recommendation execution： The role of recommendation quality［J］. Journal of Computer Information Systems， 2022， 62（6）： 1283-1296.
[4]	Choi W J， Liu Q， Shin J. Predictive analytics and ship-then-shop subscription［J］. Management Science， 2024， 70（2）： 1012-1028.
[5]	Wang J， Qi Y. Personalized recommendation research based on logistic regression algorithm for Amazon product reviews［J］. Journal of Innovations in Economics & Management， 2024， 5（1）： 141-160.
[6]	肖静华，吴瑶，刘意，等. 消费者数据化参与的研发创新——企业与消费者协同演化视角的双案例研究［J］. 管理世界， 2018， 34（8）： 154-173+192.
	Xiao J H， Wu Y， Liu Y， et al. New product development innovation with customer digitalized engagement： A comparative case study from the firm-customer coevolutionary perspective［J］. Journal of Management World， 2018， 34（8）： 154-173+192.
[7]	Gallin S， Portes A. Online shopping： How can algorithm performance expectancy enhance impulse buying ［J］. Journal of Retailing and Consumer Services， 2024， 81： 103988.
[8]	Hinz O， Eckert J. The impact of search and recommendation systems on sales in electronic commerce［J］. Business & Information Systems Engineering， 2010， 2（2）： 67-77.
[9]	Dias M B， Locher D， Li M， et al. The value of personalised recommender systems to e-business： A case study［C］//Proceedings of the 2008 ACM Conference on Recommender Systems， Lausanne， Switzerland，October 23-25， ACM， 2008： 291-294.
[10]	Dellaert B G C， Häubl G. Searching in choice mode： Consumer decision processes in product search with recommendations［J］. Journal of Marketing Research， 2012， 49（2）： 277-288.
[11]	胡春华，赵慧，童小芹，等. 推荐系统对消费者网购支出的影响研究［J］. 中国管理科学，2020，28（6）： 158-170.
	Hu C H， Zhao H， Tong X Q， et al. Research on the influence of recommendation system on consumers' online shopping expenditure［J］. Chinese Journal of Management Science， 2020， 28（6）： 158-170.
[12]	Tian B， Han R， Zhang J， et al. Pricing and insurance strategy of online retailers for the price adjustment protection policy［J］. Electronic Commerce Research， 2025， 25（3）： 1343-1374.
[13]	Xie R， Ling C， Wang Y， et al. Deep feedback network for recommendation［C］//Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence， Yokohama，Japan，January 7-15， ACM， 2021： 2519-2525.
[14]	火山引擎. 火山引擎数智平台VeDI：新一代企业级数据智能平台［EB/OL］.（2024-09-20）［2025-11-29］. .
	Engine Volcano. VeDI （volcano engine data intelligence platform）： A new-generation enterprise-level data intelligence platform［EB/OL］.（2024-09-20）［2025-11-29］. .
[15]	Bolandifar E， Chen Z， Kouvelis P， et al. Quality signaling through crowdfunding pricing［J］. Manufacturing & Service Operations Management，2023，25（2）： 668-685.
[16]	Bandyopadhyay N， Sivakumaran B， Patro S， et al. Immediate or delayed！ Whether various types of consumer sales promotions drive impulse buying： An empirical investigation［J］. Journal of Retailing and Consumer Services， 2021， 61： 102532.
[17]	Xiao B， Benbasat I. E-commerce product recommendation agents： Use， characteristics， and impact［J］. MIS Quarterly， 2007， 31（1）： 137-209.
[18]	Lu S， Xiao L， Ding M. A video-based automated recommender VAR system for garments［J］. Marketing Science， 2016， 35（3）： 484-510.
[19]	Zhong Z. Platform search design： The roles of precision and price［J］. Marketing Science， 2023， 42（2）： 293-313.
[20]	Choudhary V， Zhang Z. Product recommendation and consumer search［J］. Journal of Management Information Systems， 2023， 40（3）： 752-777.
[21]	Lv F， Li M， Guo T， et al. XDM： Improving sequential deep matching with unclicked user behaviors for recommender system［C］//Proceedings of the 27th International Conference on Database Systems for Advanced Applications， Singapore， April 11-14 ，Cham：Springer，2022： 364-376.
[22]	Zhou B， Zou T. Competing for recommendations： The strategic impact of personalized product recommendations in online marketplaces［J］. Marketing Science， 2023， 42（2）： 360-376.
[23]	Li L， Chen J， Raghunathan S. Recommender system rethink： Implications for an electronic marketplace with competing manufacturers［J］. Information Systems Research， 2018， 29（4）： 1003-1023.
[24]	Shin J， Yu J. Targeted advertising and consumer inference［J］. Marketing Science， 2021， 40（5）： 900-922.
[25]	Zhong Z. Platform search design： The roles of precision and price［J］. Marketing Science， 2023， 42（2）： 293-313.
[26]	Yu M， Debo L， Kapuscinski R. Strategic waiting for consumer-generated quality information： Dynamic pricing of new experience goods［J］. Management Science， 2015， 62（2）： 410-435.
[27]	张福利，张燕，徐小林. 基于战略顾客行为的零售商退货策略研究［J］. 管理科学学报， 2017， 20（11）： 100-113.
	Zhang F L， Zhang Y， Xu X L. Retailer’s return policy based on strategic consumer behavior［J］. Journal of Management Sciences in China，2017，20（11）： 100-113.
[28]	周尔凤，张廷龙，倪蕾，等. 竞争中的参考价格效应及承诺［J］. 中国管理科学， 2018， 26（8）： 75-85.
	Zhou E F， Zhang T L， Ni L， et al. Advertising and price decisions considering reference price effect［J］. Chinese Journal of Management Science， 2018， 26（8）： 75-85.
[29]	Chen Y H， Jiang B. Dynamic pricing and price commitment of new experience goods［J］. Production and Operations Management， 2021， 30（8）： 2752-2764.
[30]	Chellappa R K， Mukherjee R. Platform preannouncement strategies： The strategic role of information in two-sided markets competition［J］. Management Science， 2021， 67（3）： 1527-1545.
[31]	Jullien B， Pavan A. Information management and pricing in platform markets［J］. The Review of Economic Studies， 2019， 86（4）： 1666-1703.
[32]	Ebrahimi P， Khajeheian D， Soleimani M， et al. User engagement in social network platforms： What key strategic factors determine online consumer purchase behaviour ［J］.Economic Research-Ekonomska Istraživanja， 2023， 36（1）： 2106264.
[33]	DesRoches C T. Value commitment， resolute choice， and the normative foundations of behavioural welfare economics［J］. Journal of Applied Philosophy， 2020， 37（4）： 562-577.
[34]	Tereyağoğlu N， Fader P S， Veeraraghavan S. Pricing theater seats： The value of price commitment and monotone discounting［J］. Production and Operations Management， 2017， 26（6）： 1056-1075.
[35]	Bala R， Carr S. Pricing software upgrades： The role of product improvement and user costs［J］. Production and Operations Management， 2009， 18（5）： 560-580.
[36]	Li X， Cai X， Chen J. Quality and private label encroachment strategy［J］. Production and Operations Management， 2022， 31（1）： 374-390.
[37]	Baier D， Stüber E. Acceptance of recommendations to buy in online retailing ［J］. Journal of Retailing and Consumer Services， 2010， 17（3）， 173-180.
[38]	Kumar A， Hosanagar K. Measuring the value of recommendation links on product demand［J］. Information Systems Research， 2019， 30（3）： 819-838.
[39]	Shi S， Sun J， Cheng T C E. Wholesale or drop-shipping： Contract choices of the online retailer and the manufacturer in a dual-channel supply chain［J］. International Journal of Production Economics， 2020， 226： 107618.
[40]	Jean T. The Theory of Industrial Organization［M］. Cambridge， Mass. MIT Press， 1988.
[41]	Kremer M， Mantin B， Ovchinnikov A. Dynamic pricing in the presence of myopic and strategic consumers： Theory and experiment［J］. Production and Operations Management， 2017， 26（1）： 116-133.