Generative Artificial Intelligence in Business Decision-Making: Emerging Frameworks, Enterprise Applications, and Future Challenges
TL;DR
Evidence is synthesized that GenAI functions most reliably as an augmentation tool that aggregates and structures diverse inputs for human judgment, rather than as an autonomous decision-maker, and that the primary barrier to enterprise value is organizational and workflow integration rather than model capability.
Abstract
Generative artificial intelligence (GenAI) has moved from experimental novelty to a central input in organizational decision-making, with McKinsey's Q1 2026 Global AI Survey finding that 65 percent of organizations now use generative AI in at least one business function, roughly double the adoption rate reported ten months earlier, and 72 percent report at least one AI workload in production. Despite this scale of adoption, the empirical record on business value remains sharply divided. This paper synthesizes recent academic and industry evidence, drawing on 24 sources published primarily between 2023 and 2026, to examine three interlocking questions: what theoretical frameworks currently explain GenAI's role in managerial and strategic decision-making, how enterprises are applying GenAI in practice across functional areas, and what structural challenges limit the translation of GenAI adoption into measurable business value. The paper synthesizes evidence from strategic management research on AI-assisted evaluation of business alternatives, organizational theory on GenAI's emerging roles in decision processes, and empirical field studies on ambiguity handling and sycophantic behavior in AI-generated business advice, alongside a widely cited 2025 MIT study finding that 95 percent of enterprise generative AI pilots fail to deliver measurable profit-and-loss impact. Findings indicate that GenAI functions most reliably as an augmentation tool that aggregates and structures diverse inputs for human judgment, rather than as an autonomous decision-maker, that single-model evaluations of strategic alternatives are frequently inconsistent and biased while aggregated multi-model evaluations approximate expert human judgment, and that the primary barrier to enterprise value is organizational and workflow integration rather than model capability. The paper concludes with a proposed decision-integration framework and implications for executives, AI governance functions, and researchers.