Paper Type
ERF
Abstract
This paper examines how delegating tasks to AI agents reshapes developers' capacity to simultaneously pursue exploration and exploitation tasks and occupy higher-level collaborative roles. Drawing on ambidexterity theory, we argue that AI agents relax cognitive constraints that typically force developers’ trade-offs between exploration and exploitation. By offloading routine tasks, AI delegation liberates cognitive bandwidth for exploratory work. This shift also repositions developers within collaborative workflows, as freed capacity enables them to take on roles requiring greater initiative and oversight. Using a panel of 2,305 GitHub developers observed over 33 weeks spanning the introduction of GitHub Copilot Coding Agent in May 2025, we find that AI delegation significantly increases developers' engagement in exploration tasks without reducing exploitation work, improves their ambidexterity, and elevates the role level in collaborative workflows. Our findings demonstrate that AI agents function as augmentative partners that reposition developers from downstream implementers toward upstream decision-makers.
Paper Number
1401
Recommended Citation
Tang, Leyan; Zhao, Xia; and Karahanna, Elena, "More Than Productivity? AI Agents and Developer Work’s Compositional Shift Toward Ambidexterity" (2026). AMCIS 2026 Proceedings. 6.
https://aisel.aisnet.org/amcis2026/sig_osra/sig_osra/6
More Than Productivity? AI Agents and Developer Work’s Compositional Shift Toward Ambidexterity
This paper examines how delegating tasks to AI agents reshapes developers' capacity to simultaneously pursue exploration and exploitation tasks and occupy higher-level collaborative roles. Drawing on ambidexterity theory, we argue that AI agents relax cognitive constraints that typically force developers’ trade-offs between exploration and exploitation. By offloading routine tasks, AI delegation liberates cognitive bandwidth for exploratory work. This shift also repositions developers within collaborative workflows, as freed capacity enables them to take on roles requiring greater initiative and oversight. Using a panel of 2,305 GitHub developers observed over 33 weeks spanning the introduction of GitHub Copilot Coding Agent in May 2025, we find that AI delegation significantly increases developers' engagement in exploration tasks without reducing exploitation work, improves their ambidexterity, and elevates the role level in collaborative workflows. Our findings demonstrate that AI agents function as augmentative partners that reposition developers from downstream implementers toward upstream decision-makers.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIG OSRA