Abstract
Background: Competitive intelligence (CI) is crucial for strategic decision-making, but manual approaches are generally cumbersome and limited in scope. Advancements in Artificial Intelligence (AI), like deep learning and transformer models, can automate CI and extract strategic insights from unstructured organisational data. Objectives: This paper explores the use of deep learning and transformer models in CI frameworks to enhance strategic decision-making. It also addresses linked challenges, including interpretability, data quality, and ethical difficulties. Method: A systematic qualitative literature review was conducted following PRISMA guidelines. The synthesis focused on nine articles sourced from Scopus, selected through purposive sampling against inclusion criteria limiting the search to peer-reviewed articles published between 2015 and 2025. Results: Findings reveal that deep learning and transformer models greatly boost CI functions such as sentiment analysis, competitive monitoring, and strategic forecasting. However, adoption is hindered by significant computational requirements, model interpretability, and scalability constraints. Conclusion: Deep learning and transformer models enable innovative ways to CI, providing businesses with more agile, accurate, and ethical insights. Effective implementation involves transparent deliverables, personnel training and upskilling, strong data governance frameworks, and careful consideration of intended and unforeseen implications. Contribution: This research contributes a conceptual framework that will guide strategic and ethical integration. Incorporating transformer and deep learning models into CI, thereby bridges academic theory with professional practice.