Agent-Based Data Curation Practices: Customer Responses to Human versus Algorithmic Data Requesters in Established Business-to-Business Relationships
通过实地实验和在线实验，研究了在已建立的B2B关系中，客户对算法数据请求者（ADR）与人类数据请求者（HDR）在数据丰富和数据核对任务中的不同回应，发现客户对ADR在数据丰富任务中更配合，而对HDR在数据核对任务中更偏好。
With the increasing value generated through data curation and the rise of artificial intelligence (AI) agents that act as human agents, vendor companies in established business-to-business relationships increasingly delegate data curation tasks to algorithmic data requesters (ADRs) instead of human data requesters (HDRs). Using a randomized field experiment with a European pharmaceutical company and a follow-up online experiment, we show how customers respond to ADRs versus HDRs across two tasks: data enrichment (i.e., adding new information) and data reconciliation (i.e., updating existing records). For data enrichment, customers are more likely to agree and complete requests sent by ADRs because they expect lower effort. For data reconciliation requests, customers lean toward HDRs, reflecting stronger accuracy concerns. Interestingly, we observe only a marginal advantage in completion rates for HDRs. These findings advise vendors to match agents to data curation tasks; they should deploy ADRs for data enrichment to reduce customer burden, and they should use HDRs for data reconciliation to address error concerns. Relatedly, vendors should craft emails that match the messaging context; they should frame data enrichment messages around convenience and benefits, and they should frame data reconciliation messages around accuracy, auditability, and risk reduction.