This research introduces a capability-level theory for placing accountability boundaries within agentic AI systems, introducing 'accountability assets' and 'rule debt' to address how responsibility transfers as AI capabilities become modular and distributed.
Researchers propose a new theoretical framework for understanding accountability in increasingly complex AI environments. Agentic AI orchestrators facilitate the modularization of information systems, but this modularity complicates the assignment of responsibility. The paper introduces 'accountability assets'—complementary elements necessary to make AI outputs legitimate, auditable, and assignable to a responsible party. The theory explores how verification costs and responsibility transferability determine whether execution and accountability boundaries can move together. It outlines three boundary strategies (component, integrated, dual-track) and introduces the concept of 'rule debt,' which measures the governance burden that accrues when organizational rules migrate into agentic execution environments. The framework links digital innovation, transaction costs, and governance perspectives to explain the relationship between modularization and organizational disaggregation.