Employee Reskilling for AI Adoption and Organizational Performance: A Comparative Study of Infosys and TCS

Abstract

Artificial intelligence is no longer an experimental add on for the Indian information technology industry it has become a core part of how large service providers price their work, staff their projects and talk to investors. This shift has pushed employee reskilling to the centre of corporate strategy, because a workforce that cannot operate alongside AI tools quickly turns into a cost rather than an asset. The present paper examines how two of India's largest IT services companies, Infosys Limited and Tata Consultancy Services (TCS), have responded to this pressure through their reskilling programmes, and it asks whether the scale and depth of these programmes line up with measurable business outcomes. Using only secondary data drawn from annual reports, investor press releases, regulatory filings and reputable business media published between 2019 and 2026, the study builds a composite Reskilling Intensity Index (RII) from four equally weighted indicators: average learning hours per employee, the share of the workforce trained in AI, the maturity of the internal learning platform, and the visibility of leadership commitment to reskilling. Both companies score in the 'High' intensity band, with TCS marginally ahead of Infosys once the four sub-scores are combined. The paper then places this index alongside four financial and operational indicators Return on Equity (ROE), digital/AI revenue share, operating margin and revenue per employee to see whether a heavier investment in reskilling shows up in the numbers. The findings suggest a positive but uneven relationship: TCS records a noticeably higher ROE, largely on account of its capital structure and dividend policy rather than reskilling alone, while Infosys shows steadier digital revenue conversion in the years for which comparable disclosure exists. The paper closes with a discussion of what these patterns might mean for HR and finance functions inside large IT firms, and it is honest about the limits of a two-company, secondary-data design.