Strategic Control Architecture Under Algorithmic Dependency

Abstract

As organizations delegate more decision-making to algorithmic systems, the question is not whether they maintain control — it is whether they have designed a surface through which control can be exercised at all. This paper examines the architecture of strategic control under algorithmic dependency: what decision loops must remain human, where override capacity degrades silently, and how dependency on automated systems restructures the organization's own capacity to read and correct itself. The paper proposes a framework for mapping control surfaces before delegation rather than attempting to recover them after the fact.