After Agile: How AI Agents May Redefine Enterprise Project Management
Year: 2026
Type: preprint
Abstract: Artificial-intelligence agents increasingly participate in enterprise projects by producing requirements, executing tool-mediated tasks, modifying software, inspecting operational states, coordinating workflows, and reporting progress. Such participation invites the prediction that autonomous systems will displace established Agile or predictive management methods. This conceptual paper challenges that inference. It compares Scrum, Kanban, conventional project governance, AI risk management, and recent agentic project-management research, and distinguishes the completion of a delegated action from accepted delivery, observed business effects, continuity of outstanding obligations, and authorized closure. Its proposed contribution is a claim-relative qualification perspective for connecting evidence generated by multiple agents and enterprise systems to project-level decisions. Constructed adversarial cases identify situations in which existing controls are already sufficient, cases of incomplete cross-system evidence, and cases in which legitimate project closure precedes benefits realization. The analysis argues that the consequential change may concern the scale and composition of assurance rather than the replacement of project methodologies. The proposal is a bounded integration of existing practices, not a new universal framework, and it claims no executed experiment, calibrated performance gain, independent validation, or demonstrated superiority over conventional controls.
Author: Figurelli, Rogério
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236353040
