This paper presents a hybrid architecture for agent-enhanced strategic reasoning that combines heuristic extraction, semantic activation, and constructive synthesis. Drawing from sources ranging from classical military theory to modern corporate strategy, we activate and construct multiple heuristics through a process of semantic interdependence inspired by quantum cognition research. Unlike traditional decision engines that select optimal rules, our system fuses conflicting heuristics into a coherent and context-sensitive narrative based on semantic interaction modeling and rhetorical framing. We demonstrate the framework through a case study of Meta vs. FTC, and perform preliminary validation using semantic metrics. We also discuss extensions (e.g., dynamic interference coordination) to Limitations.
