Engineering Executive Authority in Deterministic AI Systems: A Governance Architecture for Capital-Sensitive Institutions

This manuscript introduces an engineering-based governance architecture for deterministic AI systems operating in capital-sensitive institutional environments. The study formalizes Executive Authority Enforcement (EAE) as a computational precondition embedded within AI decision pipelines, ensuring that no high-impact decision executes without authority-aligned validation. The proposed framework integrates capital exposure modeling, hierarchical escalation thresholds, and deterministic validation gates to eliminate governance–execution asymmetry. Through formal theorem development, bounded exposure lemmas, and stability analysis grounded in control theory and financial risk modeling, the study demonstrates how executive authority can be engineered as a structural constraint rather than treated as post-hoc oversight. The contribution of this work lies in translating board-level risk appetite and fiduciary accountability into executable system parameters, aligning institutional governance with deterministic execution logic in capital-sensitive contexts. This preprint represents a foundational contribution within the Deterministic Decision Authority (DDA) research program and is prepared for submission to a peer-reviewed journal in the field of engineering management and institutional AI governance.