The increasing complexity of construction projects requires management architectures capable of integrating heterogeneous data, computational intelligence, human decision-making, and sustainability objectives. This paper develops a conceptual socio-technical architecture for AI-enabled sustainable construction management by synthesizing the analytical, text-mining, sentiment-analysis, and optimization capabilities represented in the provided literature. The proposed architecture consists of six interacting layers: data acquisition, semantic intelligence, analytical intelligence, optimization, human–AI coordination, and sustainability-oriented decision support. Text-mining techniques provide mechanisms for extracting structured information from unstructured project communication, while sentiment and opinion-analysis approaches support the interpretation of stakeholder perceptions. Optimization-oriented studies provide a foundation for adaptive resource allocation and computational search. The methodology uses a reference-driven conceptual synthesis rather than empirical experimentation because the supplied literature does not contain construction-specific experimental datasets. The resulting framework demonstrates how heterogeneous AI functions can be organized into a scalable management architecture while preserving human oversight. Findings indicate that scalability depends not only on computational capacity but also on semantic interoperability, explainability, adaptive optimization, and organizational acceptance. The paper identifies important limitations, particularly the indirect transferability of sentiment-analysis and optimization studies to construction environments, and proposes future empirical validation through project-level datasets and longitudinal implementation studies.