AI-Driven Natural Resource Optimization and Financial Performance in the Automotive Industry: A PRISMA-Based Systematic Review of Twelve Artificial Intelligence Technique Families

Abstract

The global automobile sector generates more than USD 2.8 trillion in annual revenue and employs over 8 million people, making it one of the most capital-intensive and operationally complex industries in the world economy, with multi-tiered supply chains spanning dozens of countries. Amid the parallel transformations of electrification, digitalisation and tightening environmental regulation, artificial intelligence (AI) has emerged as a primary enabler of financial performance across the automotive value chain. This systematic review synthesises the application of twelve AI-technique families like supervised machine learning, deep learning, natural language processing, reinforcement learning, mixed-integer linear programming (MILP), time-series econometrics, computer vision, simulation and digital twins, structural/causal inference, graph and network AI, hybrid ensembles, and federated learning to automotive financial decision-making. Following PRISMA 2025 guidelines, 4,217 records were retrieved from five major databases (2012–2025), yielding 128 eligible studies after multi-stage screening and CASP-based quality appraisal, of which five landmark studies were subjected to extended critical appraisal. We cross-tabulate the twelve AI-technique families against eight financial-performance dimensions are cash-flow optimisation, cost reduction, revenue and pricing, market stability, innovation return on investment, working-capital efficiency, quality/warranty cost, and ESG compliance. The synthesised evidence shows that MILP-based production scheduling reduces late-invoiced orders by up to 76% and inventory holding costs by 20.4%; reinforcement-learning-based dynamic pricing raises revenue by 8-12%; and computer-vision-based quality inspection reduces warranty cost exposure by USD 50-200 million per OEM per year. These gains are, however, materially conditioned by data quality, implementation context, and organisational maturity. The review further identifies ten priority research gaps are most urgently, causal identification, real-time closed-loop AI-financial integration, and explainability under emerging regulatory regimes that define an academic and industrial research agenda for the coming decade.