Artificial Intelligence Adoption, Technological Efficiency, and Production Dynamics in Digital Platforms: Evidence from Indian E-Commerce

Abstract

Artificial Intelligence (AI) is becoming increasingly embedded in the operational architecture of digital platforms, particularly in e-commerce, where firms rely on algorithms to coordinate logistics, pricing, customer interaction, warehousing, and workforce allocation. Yet, higher technology expenditure does not necessarily imply greater operational efficiency. The more relevant question is whether AI has been integrated deeply enough into routine production processes to generate measurable productivity gains.This study examines the relationship between AI adoption maturity and technical productivity across Indian e-commerce platforms. Firm-level data were collected through a structured survey of 100 e-commerce enterprises operating across asset-light marketplaces, inventory-led platforms, and specialised fulfilment networks. AI adoption is measured through a five-dimensional Capability Maturity Model covering logistics optimisation, conversational Natural Language Processing (NLP), dynamic pricing, warehouse automation, and algorithmic labour management. These dimensions are combined to form a continuous AI Adoption Index ranging from 0 to 20. An Ordinary Least Squares (OLS) log-linear production model is then estimated using current operational Gross Merchandise Value (GMV) as the output measure, while controlling for capital intensity, gig-worker density, firm age, and historical productivity. The results indicate a positive and statistically significant relationship between AI maturity and platform productivity. A one-point increase in the AI Adoption Index is associated with an estimated 6.10% increase in technical productivity (β1 = 0.06095, p < 0.001). Historical productivity is also strongly associated with current output (β5 = 0.92285, p < 0.001), highlighting the importance of accumulated organisational scale and prior performance. In contrast, general IT capital intensity and gig-worker density do not exhibit statistically significant direct effects. The model explains 99.32% of the observed variation in the dependent variable (Adjusted R² = 0.9932). The findings suggest that productivity gains are associated less with technology spending in isolation and more with the extent to which AI becomes embedded in core operational processes. The study contributes to the literature on digital transformation by demonstrating how functional AI capability, rather than technology expenditure alone, can shape productivity outcomes in India's platform economy.