Abstract

Digital platforms increasingly rely on human-mediated service interactions to support customer engagement and operational outcomes. However, in labor-intensive services, the adoption of platform-provided human agents remains uneven, and the drivers of this adoption and its effects on customer and provider outcomes are not fully understood. Drawing on a large-scale dataset of platform service transactions, this study examines how human-agent deployment interacts with service characteristics and formal safeguards to influence adoption, customer retention, purchase frequency, relationship duration, and operational efficiency. Results show that labor intensity significantly increases the likelihood of human-agent adoption, but only when formal safeguards, such as insurance mechanisms, are present. Without these safeguards, adoption remains low, reflecting customer concerns about execution risk. When safeguards are introduced, human-agent adoption rises substantially, indicating that these mechanisms are most effective when aligned with task complexity and risk allocation.  

Empirical analyses further demonstrate that human-agent adoption is associated with higher customer retention, greater purchase frequency, and longer relationship duration. On the provider side, human agent involvement is linked to reduced service time, increased use of escalations that improve transparency, and higher operational efficiency. Financial analyses show measurable benefits, including lower retargeting costs and positive return on investment. These findings contribute to research on platform operations by identifying execution-stage mechanisms as key drivers of sustained engagement, particularly in services with limited observability of effort and uncertain outcomes. The results also provide practical guidance for platform managers: deploying human agents in high-risk, labor-intensive services, combined with formal safeguards, improves both customer outcomes and firm performance.  

By linking mechanisms to measurable outcomes, this study offers a framework for understanding how human interventions can reduce execution risk and sustain engagement in digital marketplaces. 

Please contact us at dip.mkt@unibocconi.it if you wish to attend.