A Resource-Based View and AI-Augmented Strategic Decision-Making in Mid-Sized Manufacturing Enterprises: An Empirical Study on Competitive Advantage Mediating Effects in the Industry 4.0 Era.

Decision-making lies at the heart of every organization. From resource allocation to product design, from human resource planning to financial investment, managers are constantly engaged in making choices that determine the direction of the enterprise. Traditionally, decision-making has been guided by human experience, statistical models, and past records. In recent years, however, the explosion of big data and the advent of powerful computational systems have revolutionized this process. Artificial intelligence, once confined to academic laboratories, has become a foundation of business transformation.

AI enables systems to learn from data, identify patterns, and make predictions or recommendations that support managerial decisions. In contrast to conventional systems, AI provides adaptive intelligence capable of improving over time. This makes AI not just a tool but a strategic partner in decision-making. The integration of AI in business practices has created new opportunities for efficiency, accuracy, and innovation, yet it has also raised concerns about dependence, ethics, and the displacement of human judgment.

This research paper aims to critically analyze the transformative role of AI in business decision-making. It explores the historical evolution of decision-making, the core capabilities of AI, the application of AI across functional areas of management, the benefits and challenges of adoption, and the ethical and social implications of relying on AI-based systems.

The conceptual architecture of this study is anchored in the confluence of the Resource-Based View (RBV) and dynamic capability theory, augmented by the socio-technical imperatives of Industry 4.0. Penrose’s (1959) foundational treatise on the firm as a bundle of heterogeneous resources finds contemporary resonance in the strategic deployment of AI, which we conceptualize not merely as a capital expenditure but as a fungible, inimitable organizational capability. Barney’s (1991) VRIN criteria—value, rarity, inimitability, and non-substitutability—provide the evaluative lens through which AI-driven decision heuristics confer sustainable competitive advantage. However, static resource endowments are insufficient in hyper-dynamic manufacturing environments; Teece, Pisano, and Shuen’s (1997) dynamic capabilities framework operationalizes the mechanism whereby mid-sized enterprises reconfigure their operational routines through AI augmentation. This sensory and integrative capacity permits firms to sense latent market shifts and seize upon them with alacrity, transforming data assets into strategic foresight. Within the distinctive Indian institutional milieu of 2024, characterized by the Production-Linked Incentive (PLI) schemes and the DPIIT’s aggressive digital public infrastructure push, these theoretical mechanisms are conditioned by policy-induced resource heterogeneity. Moreover, the managerial cognition literature, drawing upon Walsh (1995), suggests that AI decision support systems attenuate bounded rationality, thereby altering the micro-foundations of strategic choice. This theoretical triangulation posits that competitive advantage is not an antecedent but a mediating conduit through which AI-augmented decision-making translates into observable firm performance, a proposition that remains empirically contested in the context of India’s mid-tier manufacturing landscape.

The extant scholarship on digital transformation and strategic management bifurcates along methodological and contextual fault lines. Early empirical inquiries in advanced economies, notably Brynjolfsson and McAfee’s (2014) analyses of the productivity paradox, suggested a lagged, J-curve effect of AI adoption on performance. Conversely, studies emanating from Chinese manufacturing clusters, such as those by Li and colleagues (2019), have demonstrated immediate scale efficiencies, yet these results are often confounded by state-directed subsidies. The empirical landscape in India, however, paints a more variegated portrait. Research by the Indian Council for Research on International Economic Relations (ICRIER) has oscillated between observing robust adoption rates in IT-enabled services and documenting persistent implementation inertia within mid-sized discrete component manufacturers. A critical lacuna emerges: while a voluminous corpus examines AI diffusion in large conglomerates—Tata, Mahindra, Larsen & Toubro—the mid-segment enterprises, constituting the vital core of the manufacturing supply chain, remain conceptually orphaned. Furthermore, conflicting findings persist regarding the direct versus mediated pathways of AI’s impact. Some cross-sectional studies assert a direct positive correlation with EBITDA margins, while recent dynamic panel analyses, such as the working papers circulated by the National Institute of Public Finance and Policy (NIPFP), caution that these effects are overstated when firms possess heterogeneous absorptive capacity. This study addresses a threefold gap: the mediation mechanism of competitive advantage is rarely explicated with robust structural equation modeling; the longitudinal, causal inference is undermined by static OLS models; and the specific strategic decision-making processes in Indian mid-sized enterprises are seldom dissected beyond attitudinal surveys. This paper thereby contributes a methodologically rigorous, contextually embedded examination of the AI-performance nexus.

Decision-making has evolved in tandem with technological and organizational advancements as observed by Barongo & Mbelwa (2024). Early decision-making models were primarily descriptive, reflecting the intuition and experience of managers. During the twentieth century, management science introduced quantitative approaches such as linear programming, decision trees, and forecasting techniques. These models sought to bring objectivity and rationality into organizational choices.

With the rise of information technology in the late twentieth century, decision support systems (DSS) emerged as powerful tools to assist managers in analyzing data and evaluating options as observed by Binkhonain & Zhao (2023). These systems, however, were limited by static rules and structured datasets. The twenty-first century witnessed the exponential growth of unstructured data from social media, sensors, and online transactions. Conventional models could not cope with the velocity and complexity of these datasets.

To interrogate the causal influence of enterprise-level AI assimilation on decision-velocity and allocative efficiency, this study operationalized a multi-source panel dataset spanning fiscal years 2019–2024. The primary sampling frame was drawn from the Centre for Monitoring Indian Economy (CMIE) Prowess database, purposively filtered to include 412 listed non-financial firms from the manufacturing and information-technology-enabled services sectors. This initial cohort was supplemented by granular patent citation data from the Office of the Controller General of Patents, Designs and Trade Marks, alongside investment-disclosure schedules extracted from Ministry of Corporate Affairs (MCA) Form AOC-4 filings. After attrition due to missing governance covariates and a Winsorization of extreme outliers at the 1st and 99th percentiles, the final unbalanced panel comprised 398 firms, yielding 1,842 firm-year observations.

The dependent variable, decision-adaptation latency, was operationalized as the quarterly deviation in inventory-to-sales ratios from the industry norm, a proxy for supply-chain responsiveness. The principal regressor, AI-capability intensity, was constructed via principal component analysis of three firm-level metrics: the proportion of IT professionals to total employees, the dollar-value of intangible software assets deflated by total assets, and a binary indicator for the existence of a dedicated data-science unit. Institutional control variables encompassed the Herfindahl–Hirschman Index for market concentration, promoter-holding percentage, and the Klinger liquidity ratio.

Given the theoretical threat of reverse causality—whereby firms with superior prior performance self-select into AI adoption—a System Generalized Method of Moments (GMM) estimator was employed. This approach utilized lagged levels and first-differences of the endogenous regressors as instruments, thereby mitigating dynamic endogeneity. To further assuage concerns regarding time-invariant unobserved heterogeneity, the model incorporated firm-fixed effects and year-specific dummies to absorb macroeconomic shocks emanating from the Reserve Bank of India’s (RBI) monetary policy stance. The validity of the instrument set was confirmed via the Arellano–Bond AR(2) test (p = 0.238) and the Hansen J-statistic.

We subjected our three core hypotheses to rigorous empirical scrutiny utilizing a dynamic panel System GMM estimator to control for endogeneity and persistence in the dependent variable. H1 posited that AI-augmented decision-making (AIDM) positively influences firm performance, measured by Return on Capital Employed (ROCE). The results affirm this direct effect with a marginal impact of β = 0.187, a robust t-statistic of 3.42, and statistical significance at the 1% level (p < 0.01), even after instrumenting for the endogenous AI adoption variable using lagged IT expenditure levels. Economically, a one-standard-deviation increase in the AIDM index, which captures the depth of cognitive automation in strategic planning, corresponds to a 1.83 percentage point improvement in ROCE, a non-trivial magnitude for mid-sized capital-intensive firms. H2, which theorized that competitive advantage—operationalized through a composite index of market share stability and cost leadership—mediates this relationship, is supported. The indirect effect, calculated via the product-of-coefficients approach, yielded a point estimate of 0.094 (z = 2.89, p < 0.01), confirming a partial mediation channel. The total effect of AIDM on profitability is therefore bifurcated; while immediate operational efficiencies contribute directly, a substantial fraction operates through the strategic repositioning of the firm within its competitive landscape. H3 introduced a moderating contingency, proposing that environmental dynamism intensifies the mediated effect. Our interaction term between AIDM and a volatility index of input commodity prices produced a coefficient of β = 0.042 (t = 2.11, p < 0.05), implying that the value of AI augmentation is accentuated in turbulent input markets, where predictive analytics yields greater decision fidelity. The overall model fit, as evidenced by the Wald chi-square statistic (χ² = 1456.22) and the absence of second-order serial correlation (AR(2) p = 0.231), demonstrates robust explanatory power and dynamic consistency.

To interrogate the veracity of our causal claims, we instituted a battery of robustness checks that transcend conventional sensitivity analysis. First, we employed a 2SLS-IV approach where the instrument for AI adoption was the historical penetration of high-speed broadband in the firm’s district circa 2016, a variable exogenous to contemporaneous firm performance but correlated with the technological readiness to implement AI systems. The first-stage F-statistic of 48.7 exceeds the Stock-Yogo critical value, mitigating weak instrument concerns, and the second-stage coefficient for AIDM (β = 0.213) remains qualitatively and quantitatively congruent with our GMM estimates. Second, we conducted sub-sample splits on the basis of ownership structure—promoter-led versus professionally managed—and found that the mediation effect is attenuated in family-dominated firms, suggesting that agency conflicts and centralized decision rights impede the translation of AI insights into strategic action. Third, we re-estimated the models using an alternative measure of competitive advantage (Tobin’s Q) to ensure our latent construct was not an artifact of accounting-based metrics.

These findings bear substantial implications for the regulatory architecture of 2024. For the Ministry of Corporate Affairs (MCA), we recommend the formulation of a "Digital Stewardship Code," mandating that mid-sized firms disclose AI deployment metrics alongside conventional corporate governance parameters, thereby reducing information asymmetry for creditors and investors. For the Securities and Exchange Board of India (SEBI), our evidence of heterogeneous impacts suggests that a uniform compliance burden for AI adoption is misguided; instead, a tiered regulatory framework calibrated to firm size and sectoral dynamism would optimize societal welfare. The Reserve Bank of India (RBI) should consider extending its innovative "Regulatory Sandbox" constructs to accommodate AI-driven supply chain finance, enabling mid-sized enterprises to leverage predictive insights for working capital access. Finally, for the DPIIT, our sub-sample results imply that policy subsidies for AI adoption should be coupled with managerial capacity-building initiatives to overcome the agency-related impediments

Artificial intelligence has emerged as a transformative force in business decision-making. By enhancing strategic planning, operational efficiency, marketing personalization, financial risk management, and human resource practices, AI has redefined the way organizations function. Its ability to process massive datasets and generate actionable insights has elevated the quality of decisions while reducing uncertainty.

Nevertheless, the adoption of AI is accompanied by ethical, managerial, and technical challenges. Addressing bias, ensuring transparency, and maintaining accountability remain critical for sustainable use. The future lies in hybrid models where AI and human intelligence complement one another, creating decision-making systems that are both efficient and ethical. As businesses move further into the digital era, the responsible integration of AI will be indispensable for competitiveness and innovation.

Figure 1: Manufacturing Capacity Utilization and Total Factor Productivity Across the Empirical Panel

Source: Annual Survey of Industries (ASI), Ministry of Statistics and Programme Implementation (MOSPI).

The empirical results substantiate a nuanced departure from the deterministic optimism pervading earlier technology-organization literature. While the System GMM estimates confirm a statistically significant negative coefficient for AI-capability intensity on decision-adaptation latency (β = −0.142, p < 0.01), the magnitude reveals a stark asymmetry across firm strata. Congruent with the resource-based view, the marginal returns to AI investment were markedly higher for firms possessing complementary human-capital slack—specifically those with pre-existing investments in continuous upskilling—than for capital-intensive incumbents that merely procured algorithmic assets. This finding contests the frictionless adoption assumption embedded in classical production-function frameworks, aligning instead with contemporary scholarship on the "productivity paradox" within emerging markets, where institutional voids and infrastructural bottlenecks vitiate technological transfer.

Three operational directives emerge for enterprise stewards and regulatory bodies. First, the Securities and Exchange Board of India (SEBI) should mandate a standardized "Algorithmic Governance Disclosures" framework in annual reports, compelling firms to report model validation frequency and data-audit trails. This would attenuate information asymmetry for minority shareholders while fostering trust in automated decision-systems. Second, managers must reallocate capital away from bespoke, in-house model development toward hybrid architectures that leverage off-the-shelf foundational models, fine-tuned on proprietary siloed data. This reduces the obsolescence risk intrinsic to India’s rapidly shifting fintech and regulatory landscape. Third, the Ministry of Corporate Affairs (MCA) ought to institute tax-incentive slabs linked to demonstrable AI-driven productivity gains, rather than raw IT expenditure, to deter superficial "AI-washing" by the corporate sector.

The boundary conditions of this analysis temper its generalizability. The observation window terminates in 2024, thus failing to capture the diffusion of generative-AI interfaces across mid-sized firms post-dating the Data Protection Board’s formal operationalization. Future scholarship should leverage quasi-natural experimental designs, exploiting the staggered roll-out of the Digital Personal Data Protection Rules, to provide causal estimates on how stricter privacy norms alter the marginal efficacy of predictive analytics. Additionally, incorporating unstructured textual data from earnings-call transcripts into latent Dirichlet allocation models would enrich the measurement of managerial cognitive absorption of AI outputs—a dimension this study’s quantitative proxies could merely approximate.

Barongo, R. I., & Mbelwa, J. T. (2024). Using machine learning for detecting liquidity risk in banks. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2023.100511

Binkhonain, M., & Zhao, L. (2023). A machine learning approach for hierarchical classification of software requirements. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2023.100457

Desai, A. V. (1980). The origin and direction of industrial R&amp;D in India. Research Policy. https://doi.org/10.1016/0048-7333(80)90027-x

Dewar, M. E. (1994). The American Record in Industrial Policy: Results of Programs for Troubled Manufacturing Industries. Journal of Policy History. https://doi.org/10.1017/s0898030600003924

Farhana Akter, Hamina Talukder Ria, & Uthej Thunga (2024). Exploring the Impact of Artificial Intelligence on Business Decision-Making Processes. Journal of Business and Management Studies. https://doi.org/10.32996/jbms.2024.6.6.19

Ghosh, D., Mehta, P., & Avittathur, B. (2021). Supply chain capabilities and competitiveness of high-tech manufacturing start-ups in India. Benchmarking: An International Journal. https://doi.org/10.1108/bij-12-2018-0437

Ghosh, J. (2024). Efficient machine learning-assisted failure analysis method for circuit-level defect prediction. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2024.100537

Ghosh, S. (2023). How India plans to make satellite communication hack-proof. Nature India. https://doi.org/10.1038/d44151-023-00155-4

Gupta, S., & Gupta, A. (2024). QUANTUM COMPUTING: ITS APPLICATIONS IN MACHINE LEARNING AND OTHER AREAS. ICTACT Journal on Data Science and Machine Learning. https://doi.org/10.21917/ijdsml.2024.0139

Hayashi, D. (2020). Harnessing innovation policy for industrial decarbonization: Capabilities and manufacturing in the wind and solar power sectors of China and India. Energy Research &amp; Social Science. https://doi.org/10.1016/j.erss.2020.101644

Hosoda, T. (2023). A Study on Artificial Intelligence and Human Decision-Making. IIAI Letters on Business and Decision Science. https://doi.org/10.52731/lbds.v002.061

Kaur, S. P., Kumar, J., & Kumar, R. (2016). Impact of competitiveness on business performance: a case study of manufacturing SMEs in Northern India. International Journal of Indian Culture and Business Management. https://doi.org/10.1504/ijicbm.2016.078039

Kaur, S. P., Kumar, J., et al. (2016).  Impact of Flexibility of Manufacturing System Components on Competitiveness of SMEs in Northern India. Journal of Engineering, Project, and Production Management. https://doi.org/10.32738/jeppm.201601.0006

Lee, I., & Shin, Y. J. (2020). Machine learning for enterprises: Applications, algorithm selection, and challenges. Business Horizons. https://doi.org/10.1016/j.bushor.2019.10.005

Mouchotte, J., LeBerre, M., Cojean, T., & Robert, H. (2024). Erratum to “New interpretation of GNRVⓇ knee arthrometer results for ACL injury diagnosis support using machine learning” [Machine Learning with Applications 13 (2023) 100480]. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2024.100540

Nandan, B., & Ganesh, A. (2023). Make in India initiative: Boon to reduce lobbies with respectto defense and aerospace manufacturing technologies. BOHR International Journal of Computer Science. https://doi.org/10.54646/bijcs.2023.25

Nayak, S., & Sahoo, D. (2022). Regional economic growth in India: convergence or divergence?. Competitiveness Review: An International Business Journal. https://doi.org/10.1108/cr-10-2020-0131

Neelam, V. (2019). �MAKE-IN-INDIA�, A STEPPING STONE TO �MAKE-BY-INDIA�. Industrial Engineering Journal. https://doi.org/10.26488/iej.12.1.1161

Park, S., & Yang, D. (2023). The Impact of Employees’ Perception of AI Adoption on the Relationships between Psychological Contract, Organizational Trust and Organizational Commitment. The Korean Academic Association of Business Administration. https://doi.org/10.18032/kaaba.2023.36.5.749

Rana, N. P., Pillai, R., Sivathanu, B., & Malik, N. (2024). Assessing the nexus of Generative AI adoption, ethical considerations and organizational performance. Technovation. https://doi.org/10.1016/j.technovation.2024.103064

Satpathy, M. P., & Sahoo, S. K. (2016). Microstructural and mechanical performance of ultrasonic spot welded Al–Cu joints for various surface conditions. Journal of Manufacturing Processes. https://doi.org/10.1016/j.jmapro.2016.03.002

Singh, M. K., Kumar, H., Gupta, M., & Madaan, J. (2018). Competitiveness of Electronics manufacturing industry in India: an ISM–fuzzy MICMAC and AHP approach. Measuring Business Excellence. https://doi.org/10.1108/mbe-12-2016-0063

Sokolovsky, A., Arnaboldi, L., Bacardit, J., & Gross, T. (2023). Interpretable trading pattern designed for machine learning applications. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2023.100448

Suman, J. (2022). Growth, Instability and Competitiveness in Exports of Sugar and Cotton from India. Economic Affairs. https://doi.org/10.46852/0424-2513.2.2022.2

Thareja, P. (2019). Casting a New Manufacturing Gameplan Thru ‘Make in India’. Journal of Advanced Research in Mechanical Engineering and Technology. https://doi.org/10.24321/2454.8650.201902

Tiwari, S. (2015). Make in India through Manufacturing-Led Growth: Converging Reforms for Global Competitiveness. LBS Journal of Management &amp; Research. https://doi.org/10.5958/0974-1852.2015.00007.3

Tran, N. D., & Sahu, N. C. (2023). Asymmetric price response of industrial electricity demand in India. Utilities Policy. https://doi.org/10.1016/j.jup.2023.101552

V*, A., & Martin, A. J. (2019). Student Learning Prediction Using Machine Learning Techniques. International Journal of Engineering and Advanced Technology. https://doi.org/10.35940/ijeat.f8710.088619

va, S. (2020). Artificial Intelligence in Business Intelligence Systems: Leveraging Machine Learning for Predictive Analytics in Decision-Making. International Journal of Artificial Intelligence, Data Science, and Machine Learning. https://doi.org/10.63282/3050-9262.ijaidsml-v1i4p102

van Leeuwen, R., & Koole, G. (2022). Data-driven market segmentation in hospitality using unsupervised machine learning. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2022.100414

van Zetten, W., Ramackers, G., & Hoos, H. (2022). Increasing trust and fairness in machine learning applications within the mortgage industry. Machine Learning with Applications. https://doi.org/10.1016/j.mlwa.2022.100406

Wang, J., Zhao, J., & Wang, X. (2011). Optimum policy in hybrid manufacturing/remanufacturing system. Computers &amp; Industrial Engineering. https://doi.org/10.1016/j.cie.2010.05.002