Machine Learning, Diversification, and M&A Performance
Publication Date
4-1-2026
Abstract
This paper examines how machine learning and artificial intelligence (ML/AI) capability shapes the consequences of corporate boundary change through acquisition. We argue that ML/AI is best understood here as an organizational capability for processing information, evaluating opportunities, and coordinating decisions across stages of the acquisition process. Its value, however, should depend on the degree of relatedness between acquirer and target. When firms are closer in industry and operations, ML/AI can be deployed on information that is more comparable and more relevant to post-acquisition recombination; when firms are more distant, the same capability is more constrained by noise, missing context, and complexity. Using U.S. manufacturing acquisitions from 2015 to 2022, we identify firms with ML/AI capability from SEC disclosures and conference-call transcripts and estimate acquisition effects with matched controls and multiple difference-indifferences estimators. We find that acquisitions by ML/AIcapable firms are associated with stronger outcomes overall, but that the gains are concentrated in within-industry deals and in diversified deals between more closely related firms. The same pattern is strongest when both acquirer and target possess ML/AI capability before the deal. The paper contributes to research on organizational capabilities, corporate boundaries, and technological change by showing that the value of ML/AI in acquisitions depends critically on the relatedness of the context in which it is deployed.
Document Type
Article
Keywords
Machine Learning, Artificial Intelligence, Acquisitions, Relatedness, Corporate Boundaries JEL Classification: G34, O32, L25, C23
Disciplines
Strategic Management Policy
Source
SMU Cox: Strategy (Topic)
Language
English
