Abstract

This dissertation studies how digital service platforms should manage operations as algorithmic systems and AI reshape worker behavior, incentives, and consumer outcomes. Across three chapters, I show that these technologies change what platforms can measure, how workers respond, and how firms should allocate monitoring and decision rights. Chapter 1 uses a large-scale field experiment to compare algorithm-managed gig agents with traditionally supervised employees. The results document an efficiency-experience trade-off: gig agents perform better on measured operational metrics, but consumers later exhibit worse experience- related outcomes. Strategic case transfers help explain this gap, showing how narrow algorithmic measurement can redirect effort away from less-measured dimensions of service quality. Chapter 2 studies a seven-tier algorithmic compensation ladder for gig service agents. Using six pay cliffs in a stacked regression-discontinuity design, it shows that the ladder primarily operates through labor supply and that its effects are concentrated at particular cliffs rather than distributed evenly across the incentive schedule. Chapter 3 develops a principal-agent model of human–AI collaboration in e-commerce dispute resolution to ask when platforms should rely on humans, use AI as a copilot, or automate. The analysis shows that AI capability changes both accuracy and the informativeness of human effort, producing a threshold logic: weak AI favors humans, intermediate AI favors copilots, and strong AI favors automation.

Committee Chair

Fuqiang Zhang

Committee Members

Xiang Hui, Dennis Zhang; Tat Chan; Tianjun Feng

Degree

Doctor of Philosophy (PhD)

Author's Department

Supply Chain, Operations, and Technology

Author's School

Olin Business School

Document Type

Dissertation

Date of Award

5-5-2026

Language

English (en)

Download

Available for download on Tuesday, May 15, 2029

Share

COinS