“Scaling the Returns to Automation: Algorithms, Robots and Organizational Learning”

Strategy & Business Policy

Speaker: Christian Peukert 

Professor - UNIL

Conference  Jouy-en-Josas / room meeting S227

Digital automation technologies (DAT), including algorithmic optimization and advanced robotics, are increasingly adopted, yet evidence on how organizations scale their benefits remains limited. We study the operational and labor effects of DAT using unique data from a global logistics firm covering 13,000 management units in 60 countries. Along-side granular performance and operational data, we can follow 4,000 projects and deployment decisions of algorithmic optimization tools and robotic devices. Exploiting quasi-random delays in deployment timing and policy-induced variation in technology availability, we can identify the causal effects of DAT.

We find that DAT increases revenue and profit, as well as total labor demand by about 20 percent. These effects are driven by higher capital utilization and task expansion, alongside a reduction in operative managerial hours. We show that technological complementarity matters to surface these benefits. The largest gains arise when facilities deploy algorithmic tools and robotic devices jointly rather than in isolation. Crucially, we find that implementation experience is a key driver of scale. Organization-wide experience with the implementation of DAT explains between 20 and 40 percent of the observed increases in labor demand, revenue, and profit.

Together, these findings highlight the organizational challenge of scaling automation. The benefits of robots and algorithmic optimization do not emerge automatically, but increase with accumulated and well-managed implementation experience and the deliberate exploitation of complementarities. Organizations scale the returns to automation by learning how to deploy, adapt, and combine technologies across operations, rather than by expanding technology use alone.